:: Ring Ideals
:: by Jonathan Backer , Piotr Rudnicki and Christoph Schwarzweller
::
:: Received November 20, 2000
:: Copyright (c) 2000-2017 Association of Mizar Users
:: (Stowarzyszenie Uzytkownikow Mizara, Bialystok, Poland).
:: This code can be distributed under the GNU General Public Licence
:: version 3.0 or later, or the Creative Commons Attribution-ShareAlike
:: License version 3.0 or later, subject to the binding interpretation
:: detailed in file COPYING.interpretation.
:: See COPYING.GPL and COPYING.CC-BY-SA for the full text of these
:: licenses, or see http://www.gnu.org/licenses/gpl.html and
:: http://creativecommons.org/licenses/by-sa/3.0/.
environ
vocabularies NUMBERS, RLVECT_1, ALGSTR_1, XBOOLE_0, ALGSTR_0, STRUCT_0,
VECTSP_2, VECTSP_1, BINOP_1, LATTICES, ZFMISC_1, SUBSET_1, CARD_3,
FINSEQ_1, ARYTM_3, ORDINAL4, RELAT_1, CARD_FIL, SUPINF_2, TARSKI,
ARYTM_1, MESFUNC1, CARD_1, NEWTON, NAT_1, XXREAL_0, REALSET1, FUNCT_1,
FUNCT_7, GROUP_1, PARTFUN1, PRELAMB, MCART_1, SETFAM_1, RLVECT_3, GCD_1,
LATTICE3, SQUARE_1, FINSET_1, INT_3, XCMPLX_0, IDEAL_1, RECDEF_2;
notations TARSKI, XBOOLE_0, ZFMISC_1, SUBSET_1, FINSET_1, RELAT_1, FUNCT_1,
FINSEQ_1, ORDINAL1, REALSET1, INT_3, CARD_1, NUMBERS, XCMPLX_0, XXREAL_0,
NAT_1, RELSET_1, PARTFUN1, FUNCT_2, ALGSTR_1, XXREAL_2, XTUPLE_0,
MCART_1, POLYNOM1, SETFAM_1, DOMAIN_1, FINSEQ_4, NEWTON, STRUCT_0,
ALGSTR_0, RLVECT_1, VECTSP_1, GROUP_1, BINOP_1, VECTSP_2, FUNCT_7, GCD_1,
BINOM;
constructors SETFAM_1, BINOP_1, REALSET2, ALGSTR_1, GCD_1, INT_3, POLYNOM1,
BINOM, XXREAL_2, RELSET_1, FUNCT_7, FVSUM_1, FINSEQ_4, XTUPLE_0, NEWTON;
registrations XBOOLE_0, SUBSET_1, RELAT_1, FUNCT_2, FINSET_1, XXREAL_0,
XREAL_0, NAT_1, INT_1, MEMBERED, FINSEQ_1, REALSET1, STRUCT_0, VECTSP_1,
ALGSTR_1, INT_3, VALUED_0, ALGSTR_0, XXREAL_2, CARD_1, RELSET_1,
XTUPLE_0, ORDINAL1;
requirements NUMERALS, REAL, SUBSET, BOOLE, ARITHM;
definitions TARSKI, XBOOLE_0, ALGSTR_0;
equalities CARD_1, BINOP_1, REALSET1, STRUCT_0, ALGSTR_0, ORDINAL1;
expansions TARSKI, XBOOLE_0;
theorems INT_3, TARSKI, ZFMISC_1, RLVECT_1, VECTSP_1, GCD_1, FUNCT_1, FUNCT_2,
POLYNOM1, RELAT_1, FINSET_1, NAT_1, ALGSTR_1, FINSEQ_1, FINSEQ_2,
FINSEQ_3, SETFAM_1, INT_1, FINSEQ_4, FINSEQ_5, POLYNOM2, BINOM, XBOOLE_0,
XBOOLE_1, ORDINAL1, GROUP_1, XREAL_1, DOMAIN_1, XXREAL_0, PARTFUN1,
XXREAL_2, SUBSET_1, CARD_1;
schemes NAT_1, RECDEF_1, FUNCT_2, FINSEQ_2, DOMAIN_1, INT_1;
begin :: Preliminaries
registration
cluster add-associative left_zeroed right_zeroed for non empty addLoopStr;
existence
proof
set R = the non degenerated comRing;
take R;
thus thesis;
end;
end;
registration
cluster Abelian left_zeroed right_zeroed add-cancelable well-unital
add-associative associative commutative distributive for non trivial
doubleLoopStr;
existence
proof
set R = the non degenerated comRing;
take R;
thus thesis;
end;
end;
theorem Th1:
for V being add-associative left_zeroed right_zeroed non empty
addLoopStr, v,u being Element of V holds Sum <* v,u *> = v + u
proof
let V be add-associative left_zeroed right_zeroed non empty addLoopStr, v,
u be Element of V;
<* v,u *> = <* v *> ^ <* u *> by FINSEQ_1:def 9;
then Sum<* v,u *> = Sum<* v *> + Sum<* u *> by RLVECT_1:41
.= v + Sum <* u *> by BINOM:3
.= v + u by BINOM:3;
hence thesis;
end;
begin :: Ideals
definition
let L be non empty addLoopStr, F being Subset of L;
attr F is add-closed means
:Def1:
for x, y being Element of L st x in F & y in F holds x+y in F;
end;
definition
let L be non empty multMagma, F be Subset of L;
attr F is left-ideal means
:Def2:
for p, x being Element of L st x in F holds p*x in F;
attr F is right-ideal means
:Def3:
for p, x being Element of L st x in F holds x*p in F;
end;
registration
let L be non empty addLoopStr;
cluster add-closed for non empty Subset of L;
existence
proof
set M = the carrier of L;
for u being object holds u in M implies u in the carrier of L;
then reconsider M as Subset of L by TARSKI:def 3;
reconsider M as non empty Subset of L;
take M;
thus thesis;
end;
end;
registration
let L be non empty multMagma;
cluster left-ideal for non empty Subset of L;
existence
proof
set M = the carrier of L;
for u being object holds u in M implies u in the carrier of L;
then reconsider M as Subset of L by TARSKI:def 3;
reconsider M as non empty Subset of L;
take M;
thus thesis;
end;
cluster right-ideal for non empty Subset of L;
existence
proof
set M = the carrier of L;
for u being object holds u in M implies u in the carrier of L;
then reconsider M as Subset of L by TARSKI:def 3;
reconsider M as non empty Subset of L;
take M;
thus thesis;
end;
end;
registration
let L be non empty doubleLoopStr;
cluster add-closed left-ideal right-ideal for non empty Subset of L;
existence
proof
set M = the carrier of L;
for u being object holds u in M implies u in the carrier of L;
then reconsider M as Subset of L by TARSKI:def 3;
reconsider M as non empty Subset of L;
take M;
thus thesis;
end;
cluster add-closed right-ideal for non empty Subset of L;
existence
proof
set M = the carrier of L;
for u being object holds u in M implies u in the carrier of L;
then reconsider M as Subset of L by TARSKI:def 3;
reconsider M as non empty Subset of L;
take M;
thus thesis;
end;
cluster add-closed left-ideal for non empty Subset of L;
existence
proof
set M = the carrier of L;
for u being object holds u in M implies u in the carrier of L;
then reconsider M as Subset of L by TARSKI:def 3;
reconsider M as non empty Subset of L;
take M;
thus thesis;
end;
end;
registration
let R be commutative non empty multMagma;
cluster left-ideal -> right-ideal for non empty Subset of R;
coherence
proof
let I be non empty Subset of R;
assume I is left-ideal;
then for p,x being Element of R st x in I holds x*p in I;
hence thesis;
end;
cluster right-ideal -> left-ideal for non empty Subset of R;
coherence
proof
let I be non empty Subset of R;
assume I is right-ideal;
then for p,x being Element of R st x in I holds p*x in I;
hence thesis;
end;
end;
definition
let L be non empty doubleLoopStr;
mode Ideal of L is add-closed left-ideal right-ideal non empty Subset of L;
end;
definition
let L be non empty doubleLoopStr;
mode RightIdeal of L is add-closed right-ideal non empty Subset of L;
end;
definition
let L be non empty doubleLoopStr;
mode LeftIdeal of L is add-closed left-ideal non empty Subset of L;
end;
theorem Th2:
for R being right_zeroed left_add-cancelable left-distributive
non empty doubleLoopStr, I being left-ideal non empty Subset of R holds
0.R in I
proof
let R be right_zeroed left_add-cancelable left-distributive non empty
doubleLoopStr;
let I be left-ideal non empty Subset of R;
set a = the Element of I;
0.R*a in I by Def2;
hence thesis by BINOM:1;
end;
theorem Th3:
for R being left_zeroed right_add-cancelable right-distributive
non empty doubleLoopStr, I being right-ideal non empty Subset of R holds
0.R in I
proof
let R be left_zeroed right_add-cancelable right-distributive non empty
doubleLoopStr;
let I be right-ideal non empty Subset of R;
set a = the Element of I;
a*0.R in I by Def3;
hence thesis by BINOM:2;
end;
theorem Th4:
for L being right_zeroed non empty addLoopStr holds {0.L} is add-closed
proof
let L be right_zeroed non empty addLoopStr;
let x,y be Element of L;
assume x in {0.L} & y in {0.L};
then x = 0.L & y= 0.L by TARSKI:def 1;
then x + y = 0.L by RLVECT_1:def 4;
hence thesis by TARSKI:def 1;
end;
theorem Th5:
for L being left_zeroed right_add-cancelable right-distributive
non empty doubleLoopStr holds {0.L} is left-ideal
proof
let L be left_zeroed right_add-cancelable right-distributive non empty
doubleLoopStr;
let p,x be Element of L;
reconsider p9 = p as Element of L;
assume x in {0.L};
then x = 0.L by TARSKI:def 1;
then p9*x = 0.L by BINOM:2;
hence thesis by TARSKI:def 1;
end;
theorem Th6:
for L being right_zeroed left_add-cancelable left-distributive
non empty doubleLoopStr holds {0.L} is right-ideal
proof
let L be right_zeroed left_add-cancelable left-distributive non empty
doubleLoopStr;
let p,x be Element of L;
reconsider p9 = p as Element of L;
assume x in {0.L};
then x = 0.L by TARSKI:def 1;
then x*p9 = 0.L by BINOM:1;
hence thesis by TARSKI:def 1;
end;
registration
let L be right_zeroed non empty addLoopStr;
cluster {0.L} -> add-closed for Subset of L;
coherence by Th4;
end;
registration
let L be left_zeroed right_add-cancelable right-distributive non empty
doubleLoopStr;
cluster {0.L} -> left-ideal for Subset of L;
coherence by Th5;
end;
registration
let L be right_zeroed left_add-cancelable left-distributive non empty
doubleLoopStr;
cluster {0.L} -> right-ideal for Subset of L;
coherence by Th6;
end;
theorem
for L being add-associative right_zeroed right_complementable
distributive non empty doubleLoopStr holds {0.L} is Ideal of L;
theorem
for L being add-associative right_zeroed right_complementable
right-distributive non empty doubleLoopStr holds {0.L} is LeftIdeal of L;
theorem
for L being add-associative right_zeroed right_complementable
left-distributive non empty doubleLoopStr holds {0.L} is RightIdeal of L;
theorem Th10:
for L being non empty doubleLoopStr holds the carrier of L is Ideal of L
proof
let L be non empty doubleLoopStr;
the carrier of L c= the carrier of L;
then reconsider cL = the carrier of L as Subset of L;
A1: cL is left-ideal;
A2: cL is right-ideal;
cL is add-closed;
hence thesis by A1,A2;
end;
theorem Th11:
for L being non empty doubleLoopStr holds the carrier of L is LeftIdeal of L
proof
let L be non empty doubleLoopStr;
the carrier of L c= the carrier of L;
then reconsider cL = the carrier of L as Subset of L;
A1: cL is left-ideal;
cL is add-closed;
hence thesis by A1;
end;
theorem Th12:
for L being non empty doubleLoopStr holds the carrier of L is RightIdeal of L
proof
let L be non empty doubleLoopStr;
the carrier of L c= the carrier of L;
then reconsider cL = the carrier of L as Subset of L;
A1: cL is right-ideal;
cL is add-closed;
hence thesis by A1;
end;
definition
let R be left_zeroed right_zeroed add-cancelable distributive non empty
doubleLoopStr, I be Ideal of R;
redefine attr I is trivial means
I = {0.R};
compatibility
proof
now
assume I is trivial;
then consider x being object such that
A1: I = {x} by ZFMISC_1:131;
0.R in {x} by A1,Th3;
hence I = {0.R} by A1,TARSKI:def 1;
end;
hence thesis;
end;
end;
registration
let R be non trivial left_zeroed right_zeroed add-cancelable distributive
non empty doubleLoopStr;
cluster proper for Ideal of R;
existence
proof
reconsider M = {0.R} as Ideal of R;
M is proper by SUBSET_1:def 6;
hence thesis;
end;
end;
theorem Th13:
for L being add-associative right_zeroed right_complementable
left-distributive left_unital non empty doubleLoopStr, I being left-ideal
non empty Subset of L, x being Element of L st x in I holds -x in I
proof
let L be add-associative right_zeroed right_complementable left-distributive
left_unital non empty doubleLoopStr;
let I be left-ideal non empty Subset of L;
let x being Element of L;
assume x in I;
then
A1: (- 1.L)*x in I by Def2;
0. L = 0.L*x
.= (1.L + (- 1.L))*x by RLVECT_1:def 10
.= 1.L*x + (- 1.L)*x by VECTSP_1:def 3
.= x + (- 1.L)*x;
hence thesis by A1,RLVECT_1:def 10;
end;
theorem Th14:
for L being add-associative right_zeroed right_complementable
right-distributive right_unital non empty doubleLoopStr, I being right-ideal
non empty Subset of L, x being Element of L st x in I holds -x in I
proof
let L be add-associative right_zeroed right_complementable
right-distributive right_unital non empty doubleLoopStr;
let I be right-ideal non empty Subset of L;
let x being Element of L;
assume x in I;
then
A1: x*(- 1.L) in I by Def3;
0. L = x*0.L
.= x*(1.L + (- 1.L)) by RLVECT_1:def 10
.= x*1.L + x*(- 1.L) by VECTSP_1:def 2
.= x + x*(- 1.L);
hence thesis by A1,RLVECT_1:def 10;
end;
theorem
for L being add-associative right_zeroed right_complementable
left-distributive left_unital non empty doubleLoopStr, I being LeftIdeal of L
, x,y being Element of L st x in I & y in I holds x-y in I
proof
let L being add-associative right_zeroed right_complementable
left-distributive left_unital non empty doubleLoopStr;
let I being LeftIdeal of L;
let x, y being Element of L;
assume that
A1: x in I and
A2: y in I;
- y in I by A2,Th13;
hence thesis by A1,Def1;
end;
theorem
for L being add-associative right_zeroed right_complementable
right-distributive right_unital non empty doubleLoopStr, I being RightIdeal
of L, x,y being Element of L st x in I & y in I holds x-y in I
proof
let L being add-associative right_zeroed right_complementable
right-distributive right_unital non empty doubleLoopStr;
let I being RightIdeal of L;
let x, y being Element of L;
assume that
A1: x in I and
A2: y in I;
- y in I by A2,Th14;
hence thesis by A1,Def1;
end;
theorem Th17:
for R being left_zeroed right_zeroed add-cancelable
add-associative distributive non empty doubleLoopStr, I being add-closed
right-ideal non empty Subset of R, a being Element of I, n being Element of
NAT holds n*a in I
proof
let R be left_zeroed right_zeroed add-cancelable add-associative
distributive non empty doubleLoopStr, I be add-closed right-ideal non empty
Subset of R, a be Element of I, n be Element of NAT;
defpred P[Nat] means $1*a in I;
A1: for n being Nat holds P[n] implies P[n+1]
proof
let n be Nat;
A2: (n+1)*a = 1*a + n*a by BINOM:15
.= a + n*a by BINOM:13;
assume n*a in I;
hence thesis by A2,Def1;
end;
0*a = 0.R by BINOM:12;
then
A3: P[0] by Th3;
for n being Nat holds P[n] from NAT_1:sch 2(A3,A1);
hence thesis;
end;
theorem
for R being well-unital left_zeroed right_zeroed add-cancelable
associative distributive non empty doubleLoopStr, I being right-ideal non
empty Subset of R, a being Element of I, n being Element of NAT st n <> 0
holds a|^n in I
proof
let R be well-unital left_zeroed right_zeroed add-cancelable associative
distributive non empty doubleLoopStr, I be right-ideal non empty Subset of R
, a be Element of I, n be Element of NAT;
defpred P[Nat] means a|^$1 in I;
assume
A1: n <> 0;
A2: for n being Nat st 1 <= n holds P[n] implies P[n+1]
proof
let n be Nat;
assume 1 <= n;
A3: a|^(n+1) = (a|^n)*(a|^1) by BINOM:10;
assume a|^n in I;
hence thesis by A3,Def3;
end;
a|^1 = a by BINOM:8;
then
A4: P[1];
for n being Nat st 1 <= n holds P[n] from NAT_1:sch 8(A4,A2);
hence thesis by A1,NAT_1:14;
end;
definition
let R be non empty addLoopStr, I be add-closed non empty Subset of R;
func add|(I,R) -> BinOp of I equals
(the addF of R)||I;
coherence
proof
reconsider f = (the addF of R)||I as Function of [:I,I:],the carrier of R
by FUNCT_2:32;
A1: dom f = [:I,I:] by FUNCT_2:def 1;
for x being object st x in [:I,I:] holds f.x in I
proof
let x be object;
assume
A2: x in [:I,I:];
then consider u,v being object such that
A3: u in I & v in I and
A4: x = [u,v] by ZFMISC_1:def 2;
reconsider u,v as Element of R by A3;
reconsider u,v as Element of R;
f.x = u + v by A1,A2,A4,FUNCT_1:47;
hence thesis by A3,Def1;
end;
hence thesis by A1,FUNCT_2:3;
end;
end;
definition
let R be non empty multMagma, I be right-ideal non empty Subset of R;
func mult|(I,R) -> BinOp of I equals
(the multF of R)||I;
coherence
proof
reconsider f = (the multF of R)||I as Function of [:I,I:],the carrier of R
by FUNCT_2:32;
A1: dom f = [:I,I:] by FUNCT_2:def 1;
for x being object st x in [:I,I:] holds f.x in I
proof
let x be object;
assume x in [:I,I:];
then consider u,v being object such that
A2: u in I & v in I and
A3: x = [u,v] by ZFMISC_1:def 2;
reconsider u,v as Element of I by A2;
f.x = (the multF of R).[u,v] by A1,A3,FUNCT_1:47
.= u*v;
hence thesis by Def3;
end;
hence thesis by A1,FUNCT_2:3;
end;
end;
definition
let R be non empty addLoopStr, I be add-closed non empty Subset of R;
func Gr(I,R) -> non empty addLoopStr equals
addLoopStr (#I,add|(I,R),In (0.R,I)#);
coherence;
end;
registration
let R be left_zeroed right_zeroed add-cancelable add-associative
distributive non empty doubleLoopStr, I be add-closed non empty Subset of R;
cluster Gr(I,R) -> add-associative;
coherence
proof
set M = addLoopStr (#I,add|(I,R),In (0.R,I)#);
reconsider M as non empty addLoopStr;
now
let u be object;
A1: dom(the addF of R) = [:the carrier of R,the carrier of R:] by
FUNCT_2:def 1;
assume u in [:I,I:];
hence u in dom(the addF of R) by A1;
end;
then [:I,I:] c= dom(the addF of R);
then
A2: dom((the addF of R)||I) = [:I,I:] by RELAT_1:62;
A3: for a,b being Element of M, a9,b9 being Element of I st a9 = a & b9 =
b holds a + b = a9 + b9
proof
let a,b be Element of M, a9,b9 be Element of I;
assume
A4: a9 = a & b9 = b;
[a9,b9] in dom((the addF of R)||I) by A2;
hence thesis by A4,FUNCT_1:47;
end;
now
let a,b,c be Element of M;
reconsider a9 = a, b9 = b, c9 = c as Element of I;
a9 + b9 in I by Def1;
then
A5: [a9+b9,c9] in dom((the addF of R)||I) by A2,ZFMISC_1:def 2;
b9 + c9 in I by Def1;
then
A6: [a9,b9+c9] in dom((the addF of R)||I) by A2,ZFMISC_1:def 2;
thus (a + b) + c = ((the addF of R)||I).[a9+b9,c9] by A3
.= (a9+b9) + c9 by A5,FUNCT_1:47
.= a9 + (b9 + c9) by RLVECT_1:def 3
.= (add|(I,R)).[a9,b9+c9] by A6,FUNCT_1:47
.= a + (b + c) by A3;
end;
hence thesis by RLVECT_1:def 3;
end;
end;
registration
let R be left_zeroed right_zeroed add-cancelable add-associative
distributive non empty doubleLoopStr, I be add-closed right-ideal non empty
Subset of R;
cluster Gr(I,R) -> right_zeroed;
coherence
proof
set M = addLoopStr (#I,add|(I,R),In (0.R,I)#);
reconsider M as non empty addLoopStr;
now
let u be object;
A1: dom(the addF of R) = [:the carrier of R,the carrier of R:] by
FUNCT_2:def 1;
assume u in [:I,I:];
hence u in dom(the addF of R) by A1;
end;
then [:I,I:] c= dom(the addF of R);
then
A2: dom((the addF of R)||I) = [:I,I:] by RELAT_1:62;
now
let a be Element of M;
reconsider a9 = a as Element of I;
0.R in I by Th3;
then
A3: [a9,0.R] in dom((the addF of R)||I) by A2,ZFMISC_1:def 2;
thus a + 0.M = ((the addF of R)||I).[a9,0.R] by Th3,SUBSET_1:def 8
.= a9 + 0.R by A3,FUNCT_1:47
.= a by RLVECT_1:def 4;
end;
hence thesis by RLVECT_1:def 4;
end;
end;
registration
let R be Abelian non empty doubleLoopStr, I be add-closed non empty
Subset of R;
cluster Gr(I,R) -> Abelian;
coherence
proof
set M = addLoopStr (#I,add|(I,R),In (0.R,I)#);
reconsider M as non empty addLoopStr;
now
let u be object;
A1: dom(the addF of R) = [:the carrier of R,the carrier of R:] by
FUNCT_2:def 1;
assume u in [:I,I:];
hence u in dom(the addF of R) by A1;
end;
then [:I,I:] c= dom(the addF of R);
then
A2: dom((the addF of R)||I) = [:I,I:] by RELAT_1:62;
A3: for a,b being Element of M, a9,b9 being Element of I st a9 = a & b9 =
b holds a + b = a9 + b9
proof
let a,b be Element of M, a9,b9 be Element of I;
assume
A4: a9 = a & b9 = b;
[a9,b9] in dom((the addF of R)||I) by A2;
hence thesis by A4,FUNCT_1:47;
end;
now
let a,b be Element of M;
reconsider a9 = a, b9 = b as Element of I;
thus a + b = a9 + b9 by A3
.= b + a by A3;
end;
hence thesis by RLVECT_1:def 2;
end;
end;
registration
let R be Abelian right_unital left_zeroed right_zeroed right_complementable
add-associative distributive non empty doubleLoopStr, I be add-closed
right-ideal non empty Subset of R;
cluster Gr(I,R) -> right_complementable;
coherence
proof
set M = addLoopStr (#I,add|(I,R),In (0.R,I)#);
reconsider M as non empty addLoopStr;
now
let u be object;
A1: dom(the addF of R) = [:the carrier of R,the carrier of R:] by
FUNCT_2:def 1;
assume u in [:I,I:];
hence u in dom(the addF of R) by A1;
end;
then [:I,I:] c= dom(the addF of R);
then
A2: dom((the addF of R)||I) = [:I,I:] by RELAT_1:62;
A3: for a,b being Element of M, a9,b9 being Element of I st a9 = a & b9 =
b holds a + b = a9 + b9
proof
let a,b be Element of M, a9,b9 be Element of I;
assume
A4: a9 = a & b9 = b;
[a9,b9] in dom((the addF of R)||I) by A2;
hence thesis by A4,FUNCT_1:47;
end;
reconsider I as RightIdeal of R;
M is right_complementable
proof
let a be Element of M;
reconsider a9 = a as Element of I;
reconsider b = -a9 as Element of M by Th14;
a + b = a9 + -a9 by A3
.= 0.R by RLVECT_1:5
.= 0.M by Th3,SUBSET_1:def 8;
hence ex b being Element of M st a + b = 0.M;
end;
hence thesis;
end;
end;
Lm1: for R being comRing, a being Element of R holds the set of all
a*r where r is Element of R is Ideal of R
proof
let R be comRing, a be Element of R;
set M = the set of all a*r where r is Element of R ;
A1: now
let u be object;
assume u in M;
then ex r being Element of R st u = a*r;
hence u in the carrier of R;
end;
a*1.R in M;
then reconsider M as non empty Subset of R by A1,TARSKI:def 3;
reconsider M as non empty Subset of R;
A2: now
let b,c be Element of R;
assume that
A3: b in M and
A4: c in M;
consider r being Element of R such that
A5: a*r = b by A3;
consider s being Element of R such that
A6: a*s = c by A4;
b + c = a*(r + s) by A5,A6,VECTSP_1:def 7;
hence b + c in M;
end;
A7: now
let s,b be Element of R;
assume b in M;
then consider r being Element of R such that
A8: a*r = b;
s*b = (s*r)*a by A8,GROUP_1:def 3;
hence s*b in M;
end;
now
let s,b be Element of R;
assume b in M;
then consider r being Element of R such that
A9: a*r = b;
b*s = a*(r*s) by A9,GROUP_1:def 3;
hence b*s in M;
end;
hence thesis by A2,A7,Def1,Def2,Def3;
end;
theorem Th19:
for R being right_unital non empty doubleLoopStr, I being
left-ideal non empty Subset of R holds I is proper iff not(1.R in I)
proof
let R be right_unital non empty doubleLoopStr, I be left-ideal non empty
Subset of R;
A1: now
assume
A2: I is proper;
thus not 1.R in I
proof
assume
A3: 1.R in I;
A4: now
let u be object;
assume u in the carrier of R;
then reconsider u9 = u as Element of R;
u9*1.R = u9;
hence u in I by A3,Def2;
end;
for u being object holds u in I implies u in the carrier of R;
then I = the carrier of R by A4,TARSKI:2;
hence thesis by A2,SUBSET_1:def 6;
end;
end;
now
assume not 1.R in I;
then I <> the carrier of R;
hence I is proper by SUBSET_1:def 6;
end;
hence thesis by A1;
end;
theorem
for R being left_unital right_unital non empty doubleLoopStr, I
being right-ideal non empty Subset of R holds I is proper iff for u being
Element of R st u is unital holds not(u in I)
proof
let R be left_unital right_unital non empty doubleLoopStr, I be
right-ideal non empty Subset of R;
A1: now
assume
A2: I is proper;
A3: not 1.R in I
proof
assume
A4: 1.R in I;
A5: now
let u be object;
assume u in the carrier of R;
then reconsider u9 = u as Element of R;
1.R*u9 = u9;
hence u in I by A4,Def3;
end;
for u being object holds u in I implies u in the carrier of R;
then I = the carrier of R by A5,TARSKI:2;
hence thesis by A2,SUBSET_1:def 6;
end;
thus for u being Element of R st u is unital holds not u in I
proof
let u be Element of R;
assume u is unital;
then u divides 1.R by GCD_1:def 2;
then ex b being Element of R st 1.R = u*b by GCD_1:def 1;
hence thesis by A3,Def3;
end;
end;
now
1.R divides 1.R;
then
A6: 1.R is unital by GCD_1:def 2;
assume for u being Element of R st u is unital holds not u in I;
then I <> the carrier of R by A6;
hence I is proper by SUBSET_1:def 6;
end;
hence thesis by A1;
end;
theorem
for R being right_unital non empty doubleLoopStr, I being left-ideal
right-ideal non empty Subset of R holds I is proper iff for u being Element
of R st u is unital holds not(u in I)
proof
let R be right_unital non empty doubleLoopStr, I be left-ideal right-ideal
non empty Subset of R;
A1: now
assume
A2: I is proper;
A3: not 1.R in I
proof
assume
A4: 1.R in I;
A5: now
let u be object;
assume u in the carrier of R;
then reconsider u9 = u as Element of R;
u9*1.R = u9;
hence u in I by A4,Def2;
end;
for u being object holds u in I implies u in the carrier of R;
then I = the carrier of R by A5,TARSKI:2;
hence thesis by A2,SUBSET_1:def 6;
end;
thus for u being Element of R st u is unital holds not u in I
proof
let u be Element of R;
assume u is unital;
then u divides 1.R by GCD_1:def 2;
then ex b being Element of R st 1.R = u*b by GCD_1:def 1;
hence thesis by A3,Def3;
end;
end;
now
1.R divides 1.R;
then
A6: 1.R is unital by GCD_1:def 2;
assume for u being Element of R st u is unital holds not u in I;
then I <> the carrier of R by A6;
hence I is proper by SUBSET_1:def 6;
end;
hence thesis by A1;
end;
theorem
for R being non degenerated comRing holds R is Field iff for I being
Ideal of R holds (I = {0.R} or I = the carrier of R)
proof
let R be non degenerated comRing;
A1: now
assume
A2: R is Field;
thus for I being Ideal of R holds (I = {0.R} or I = the carrier of R)
proof
let I be Ideal of R;
assume
A3: I <> {0.R};
reconsider R as Field by A2;
ex a being Element of R st a in I & a <> 0.R
proof
assume
A4: not(ex a being Element of R st a in I & a <> 0.R);
A5: now
let u be object;
assume u in I;
then reconsider u9 = u as Element of I;
u9 = 0.R by A4;
hence u in {0.R} by TARSKI:def 1;
end;
now
let u be object;
assume
A6: u in {0.R};
then reconsider u9 = u as Element of R;
u9 = 0.R by A6,TARSKI:def 1;
hence u in I by Th3;
end;
hence thesis by A3,A5,TARSKI:2;
end;
then consider a being Element of R such that
A7: a in I and
A8: a <> 0.R;
ex b being Element of R st b*a = 1.R by A8,VECTSP_1:def 9;
then 1.R in I by A7,Def3;
then I is non proper by Th19;
hence thesis by SUBSET_1:def 6;
end;
end;
now
assume
A9: for I being Ideal of R holds (I= {0.R} or I = the carrier of R );
now
let a be Element of R;
reconsider a9 = a as Element of R;
reconsider M = the set of all a9*r where r is Element of R as
Ideal of R by Lm1;
a*1.R = a;
then
A10: a in M;
assume a <> 0.R;
then M <> {0.R} by A10,TARSKI:def 1;
then M = the carrier of R by A9;
then 1.R in M;
then ex b being Element of R st a*b = 1.R;
hence ex b being Element of R st b*a = 1.R;
end;
hence R is Field by VECTSP_1:def 9;
end;
hence thesis by A1;
end;
begin :: Linear combinations
definition
let R be non empty multLoopStr, A be non empty Subset of R;
mode LinearCombination of A -> FinSequence of the carrier of R means
:Def8:
for i being set st i in dom it
ex u,v being Element of R, a being Element of A st it/.i = u*a*v;
existence
proof
set p = <*>(the carrier of R);
take p;
let i be set;
thus thesis;
end;
mode LeftLinearCombination of A -> FinSequence of the carrier of R means
:Def9:
for i being set st i in dom it
ex u being Element of R, a being Element of A st it/.i = u*a;
existence
proof
set a = the Element of A;
reconsider aR = a as Element of R;
reconsider a9 = a*a as Element of R;
set p = <*a9*>;
take p;
let i be set;
assume
A1: i in dom p;
take aR, a;
dom p = {1} by FINSEQ_1:2,38;
then
A2: i = 1 by A1,TARSKI:def 1;
thus p/.i = p.i by A1,PARTFUN1:def 6
.= aR*a by A2,FINSEQ_1:40;
end;
mode RightLinearCombination of A -> FinSequence of the carrier of R means
:Def10:
for i being set st i in dom it
ex u being Element of R, a being Element of A st it/.i = a*u;
existence
proof
set a = the Element of A;
reconsider aR = a as Element of R;
reconsider a9 = a*a as Element of R;
set p = <*a9*>;
take p;
let i be set;
assume
A3: i in dom p;
take aR, a;
dom p = {1} by FINSEQ_1:2,38;
then
A4: i = 1 by A3,TARSKI:def 1;
thus p/.i = p.i by A3,PARTFUN1:def 6
.= a*aR by A4,FINSEQ_1:40;
end;
end;
registration
let R be non empty multLoopStr, A be non empty Subset of R;
cluster non empty for LinearCombination of A;
existence
proof
set a = the Element of A;
set u = the Element of R;
set v=u;
reconsider p = <*u*a*v*> as FinSequence of the carrier of R;
take p;
now
let i be set;
assume
A1: i in dom p;
take u,v, a;
i in {1} by A1,FINSEQ_1:2,38;
then i = 1 by TARSKI:def 1;
hence p/.i = u*a*v by FINSEQ_4:16;
end;
hence thesis by Def8;
end;
cluster non empty for LeftLinearCombination of A;
existence
proof
set a = the Element of A;
set u = the Element of R;
reconsider p = <*u*a*> as FinSequence of the carrier of R;
take p;
now
let i be set;
assume
A2: i in dom p;
take u, a;
i in {1} by A2,FINSEQ_1:2,38;
then i = 1 by TARSKI:def 1;
hence p/.i = u*a by FINSEQ_4:16;
end;
hence thesis by Def9;
end;
cluster non empty for RightLinearCombination of A;
existence
proof
set a = the Element of A;
set v = the Element of R;
reconsider p = <*a*v*> as FinSequence of the carrier of R;
take p;
now
let i be set;
assume
A3: i in dom p;
take v, a;
i in {1} by A3,FINSEQ_1:2,38;
then i = 1 by TARSKI:def 1;
hence p/.i = a*v by FINSEQ_4:16;
end;
hence thesis by Def10;
end;
end;
definition
let R be non empty multLoopStr, A,B be non empty Subset of R, F be
LinearCombination of A, G be LinearCombination of B;
redefine func F^G -> LinearCombination of (A \/ B);
coherence
proof
set H = F^G;
thus H is LinearCombination of (A \/ B)
proof
let i be set;
assume
A1: i in dom H;
then reconsider i as Element of NAT;
per cases;
suppose
A2: i in dom F;
then
A3: F/.i = F.i & F.i = H.i by FINSEQ_1:def 7,PARTFUN1:def 6;
consider u,v being Element of R, a being Element of A such that
A4: F/.i = u*a*v by A2,Def8;
a in A \/ B by XBOOLE_0:def 3;
hence thesis by A1,A4,A3,PARTFUN1:def 6;
end;
suppose
not i in dom F;
then consider n being Nat such that
A5: n in dom G and
A6: i=len F + n by A1,FINSEQ_1:25;
A7: G/.n = G.n & G.n = H.i by A5,A6,FINSEQ_1:def 7,PARTFUN1:def 6;
consider u,v being Element of R, b being Element of B such that
A8: G/.n = u*b*v by A5,Def8;
b in A \/ B by XBOOLE_0:def 3;
hence thesis by A1,A8,A7,PARTFUN1:def 6;
end;
end;
end;
end;
theorem Th23:
for R be associative non empty multLoopStr, A be non empty
Subset of R, a be Element of R, F be LinearCombination of A holds a*F is
LinearCombination of A
proof
let R be associative non empty multLoopStr, A be non empty Subset of R, a
be Element of R, F be LinearCombination of A;
let i be set;
assume i in dom (a*F);
then
A1: i in dom F by POLYNOM1:def 1;
then consider u, v being Element of R, b being Element of A such that
A2: F/.i = u*b*v by Def8;
take x = a*u, v, b;
thus (a*F)/.i = a*F/.i by A1,POLYNOM1:def 1
.= a*(u*b)*v by A2,GROUP_1:def 3
.= x*b*v by GROUP_1:def 3;
end;
theorem Th24:
for R be associative non empty multLoopStr, A be non empty
Subset of R, a be Element of R, F be LinearCombination of A holds F*a is
LinearCombination of A
proof
let R be associative non empty multLoopStr, A be non empty Subset of R, a
be Element of R, F be LinearCombination of A;
let i be set;
assume i in dom (F*a);
then
A1: i in dom F by POLYNOM1:def 2;
then consider u, v being Element of R, b being Element of A such that
A2: F/.i = u*b*v by Def8;
take u, x = v*a, b;
thus (F*a)/.i = (F/.i)*a by A1,POLYNOM1:def 2
.= u*(b*v)*a by A2,GROUP_1:def 3
.= u*(b*v*a) by GROUP_1:def 3
.= u*(b*(v*a)) by GROUP_1:def 3
.= u*b*x by GROUP_1:def 3;
end;
theorem Th25:
for R being right_unital non empty multLoopStr, A being non
empty Subset of R, f being LeftLinearCombination of A holds f is
LinearCombination of A
proof
let R be right_unital non empty multLoopStr, A be non empty Subset of R, f
be LeftLinearCombination of A;
let i be set;
assume i in dom f;
then consider r being Element of R, a being Element of A such that
A1: f/.i = r*a by Def9;
f/.i = r*a*1.R by A1;
hence thesis;
end;
definition
let R be non empty multLoopStr, A,B be non empty Subset of R, F be
LeftLinearCombination of A, G be LeftLinearCombination of B;
redefine func F^G -> LeftLinearCombination of (A \/ B);
coherence
proof
set H = F^G;
thus H is LeftLinearCombination of (A \/ B)
proof
let i be set;
assume
A1: i in dom H;
then reconsider i as Element of NAT;
per cases;
suppose
A2: i in dom F;
then
A3: F/.i = F.i & F.i = H.i by FINSEQ_1:def 7,PARTFUN1:def 6;
consider u being Element of R, a being Element of A such that
A4: F/.i = u*a by A2,Def9;
a in A \/ B by XBOOLE_0:def 3;
hence thesis by A1,A4,A3,PARTFUN1:def 6;
end;
suppose
not i in dom F;
then consider n being Nat such that
A5: n in dom G and
A6: i=len F + n by A1,FINSEQ_1:25;
A7: G/.n = G.n & G.n = H.i by A5,A6,FINSEQ_1:def 7,PARTFUN1:def 6;
consider u being Element of R, b being Element of B such that
A8: G/.n = u*b by A5,Def9;
b in A \/ B by XBOOLE_0:def 3;
hence thesis by A1,A8,A7,PARTFUN1:def 6;
end;
end;
end;
end;
theorem Th26:
for R be associative non empty multLoopStr, A be non empty
Subset of R, a be Element of R, F be LeftLinearCombination of A holds
a*F is LeftLinearCombination of A
proof
let R be associative non empty multLoopStr, A be non empty Subset of R, a
be Element of R, F be LeftLinearCombination of A;
let i be set;
assume i in dom (a*F);
then
A1: i in dom F by POLYNOM1:def 1;
then consider u being Element of R, b being Element of A such that
A2: F/.i = u*b by Def9;
take x = a*u, b;
thus (a*F)/.i = a*F/.i by A1,POLYNOM1:def 1
.= x*b by A2,GROUP_1:def 3;
end;
theorem
for R be non empty multLoopStr, A be non empty Subset of R,
a be Element of R, F be LeftLinearCombination of A holds
F*a is LinearCombination of A
proof
let R be non empty multLoopStr, A be non empty Subset of R, a be Element of
R, F be LeftLinearCombination of A;
let i be set;
reconsider c = a as Element of R;
assume i in dom (F*a);
then
A1: i in dom F by POLYNOM1:def 2;
then consider u being Element of R, b being Element of A such that
A2: F/.i = u*b by Def9;
take u, c, b;
thus thesis by A1,A2,POLYNOM1:def 2;
end;
theorem Th28:
for R being left_unital non empty multLoopStr, A being non
empty Subset of R, f being RightLinearCombination of A holds
f is LinearCombination of A
proof
let R be left_unital non empty multLoopStr, A be non empty Subset of R, f
be RightLinearCombination of A;
let i be set;
assume i in dom f;
then consider r being Element of R, a being Element of A such that
A1: f/.i = a*r by Def10;
f/.i = 1.R*a*r by A1;
hence thesis;
end;
definition
let R be non empty multLoopStr, A,B be non empty Subset of R, F be
RightLinearCombination of A, G be RightLinearCombination of B;
redefine func F^G -> RightLinearCombination of (A \/ B);
coherence
proof
set H = F^G;
thus H is RightLinearCombination of (A \/ B)
proof
let i be set;
assume
A1: i in dom H;
then reconsider i as Element of NAT;
per cases;
suppose
A2: i in dom F;
then
A3: F/.i = F.i & F.i = H.i by FINSEQ_1:def 7,PARTFUN1:def 6;
consider u being Element of R, a being Element of A such that
A4: F/.i = a*u by A2,Def10;
a in A \/ B by XBOOLE_0:def 3;
hence thesis by A1,A4,A3,PARTFUN1:def 6;
end;
suppose
not i in dom F;
then consider n being Nat such that
A5: n in dom G and
A6: i=len F + n by A1,FINSEQ_1:25;
A7: G/.n = G.n & G.n = H.i by A5,A6,FINSEQ_1:def 7,PARTFUN1:def 6;
consider u being Element of R, b being Element of B such that
A8: G/.n = b*u by A5,Def10;
b in A \/ B by XBOOLE_0:def 3;
hence thesis by A1,A8,A7,PARTFUN1:def 6;
end;
end;
end;
end;
theorem Th29:
for R be associative non empty multLoopStr, A be non empty
Subset of R, a be Element of R, F be RightLinearCombination of A holds
F*a is RightLinearCombination of A
proof
let R be associative non empty multLoopStr, A be non empty Subset of R, a
be Element of R, F be RightLinearCombination of A;
let i be set;
assume i in dom (F*a);
then
A1: i in dom F by POLYNOM1:def 2;
then consider u being Element of R, b being Element of A such that
A2: F/.i = b*u by Def10;
take x = u*a, b;
thus (F*a)/.i=(F/.i)*a by A1,POLYNOM1:def 2
.= b*x by A2,GROUP_1:def 3;
end;
theorem
for R be associative non empty multLoopStr, A be non empty Subset of
R, a be Element of R, F be RightLinearCombination of A holds
a*F is LinearCombination of A
proof
let R be associative non empty multLoopStr, A be non empty Subset of R, a
be Element of R, F be RightLinearCombination of A;
let i be set;
reconsider c = a as Element of R;
assume i in dom (a*F);
then
A1: i in dom F by POLYNOM1:def 1;
then consider u being Element of R, b being Element of A such that
A2: F/.i = b*u by Def10;
take c, u, b;
thus (a*F)/.i=a*(F/.i) by A1,POLYNOM1:def 1
.= c*b*u by A2,GROUP_1:def 3;
end;
theorem Th31:
for R being commutative associative non empty multLoopStr, A
being non empty Subset of R, f being LinearCombination of A holds
f is LeftLinearCombination of A & f is RightLinearCombination of A
proof
let R being commutative associative non empty multLoopStr, A being non
empty Subset of R, f being LinearCombination of A;
hereby
let i be set;
assume i in dom f;
then consider r,s being Element of R, a being Element of A such that
A1: f/.i = r*a*s by Def8;
f/.i = (r*s)*a by A1,GROUP_1:def 3;
hence ex r being Element of R, a being Element of A st f/.i = r*a;
end;
let i be set;
assume i in dom f;
then consider r,s being Element of R, a being Element of A such that
A2: f/.i = r*a*s by Def8;
f/.i = a*(r*s) by A2,GROUP_1:def 3;
hence ex r being Element of R, a being Element of A st f/.i = a*r;
end;
theorem Th32:
for S being non empty doubleLoopStr, F being non empty Subset of
S, lc being non empty LinearCombination of F ex p being LinearCombination of F,
e being Element of S st lc = p^<* e *> & <*e*> is LinearCombination of F
proof
let S be non empty doubleLoopStr, F be non empty Subset of S, lc be non
empty LinearCombination of F;
len lc <> 0;
then consider
p being FinSequence of the carrier of S, e being Element of S such
that
A1: lc = p^<*e*> by FINSEQ_2:19;
now
let i be set;
assume
A2: i in dom p;
then reconsider i1=i as Element of NAT;
A3: dom p c= dom lc by A1,FINSEQ_1:26;
then consider u, v being Element of S, a being Element of F such that
A4: lc/.i = u*a*v by A2,Def8;
take u, v, a;
thus p/.i = p.i by A2,PARTFUN1:def 6
.= lc.i1 by A1,A2,FINSEQ_1:def 7
.= u*a*v by A2,A3,A4,PARTFUN1:def 6;
end;
then reconsider p as LinearCombination of F by Def8;
A5: len lc = len p +1 by A1,FINSEQ_2:16;
take p;
take e;
thus lc = p^<* e *> by A1;
let i be set such that
A6: i in dom <*e*>;
A7: len lc in dom lc by FINSEQ_5:6;
then
A8: lc/.(len lc) = lc.(len lc) by PARTFUN1:def 6;
dom <*e*> = {1} by FINSEQ_1:2,38;
then
A9: i = 1 by A6,TARSKI:def 1;
consider u, v being Element of S, a being Element of F such that
A10: lc/.(len lc) = u*a*v by A7,Def8;
take u, v, a;
thus <*e*>/.i = <*e*>.i by A6,PARTFUN1:def 6
.= e by A9,FINSEQ_1:40
.= u*a*v by A1,A5,A10,A8,FINSEQ_1:42;
end;
theorem Th33:
for S being non empty doubleLoopStr, F being non empty Subset of
S, lc being non empty LeftLinearCombination of F ex p being
LeftLinearCombination of F, e being Element of S st lc = p^<* e *> & <*e*> is
LeftLinearCombination of F
proof
let S be non empty doubleLoopStr, F be non empty Subset of S, lc be non
empty LeftLinearCombination of F;
len lc <> 0;
then consider
p being FinSequence of the carrier of S, e being Element of S such
that
A1: lc = p^<*e*> by FINSEQ_2:19;
now
let i be set;
assume
A2: i in dom p;
then reconsider i1=i as Element of NAT;
A3: dom p c= dom lc by A1,FINSEQ_1:26;
then consider u being Element of S, a being Element of F such that
A4: lc/.i = u*a by A2,Def9;
take u, a;
thus p/.i = p.i by A2,PARTFUN1:def 6
.= lc.i1 by A1,A2,FINSEQ_1:def 7
.= u*a by A2,A3,A4,PARTFUN1:def 6;
end;
then reconsider p as LeftLinearCombination of F by Def9;
A5: len lc = len p +1 by A1,FINSEQ_2:16;
take p;
take e;
thus lc = p^<* e *> by A1;
let i be set such that
A6: i in dom <*e*>;
A7: len lc in dom lc by FINSEQ_5:6;
then
A8: lc/.(len lc) = lc.(len lc) by PARTFUN1:def 6;
dom <*e*> = {1} by FINSEQ_1:2,38;
then
A9: i = 1 by A6,TARSKI:def 1;
consider u being Element of S, a being Element of F such that
A10: lc/.(len lc) = u*a by A7,Def9;
take u, a;
thus <*e*>/.i = <*e*>.i by A6,PARTFUN1:def 6
.= e by A9,FINSEQ_1:40
.= u*a by A1,A5,A10,A8,FINSEQ_1:42;
end;
theorem Th34:
for S being non empty doubleLoopStr, F being non empty Subset of
S, lc being non empty RightLinearCombination of F ex p being
RightLinearCombination of F, e being Element of S st lc = p^<* e *> & <*e*> is
RightLinearCombination of F
proof
let S be non empty doubleLoopStr, F be non empty Subset of S, lc be non
empty RightLinearCombination of F;
len lc <> 0;
then consider
p being FinSequence of the carrier of S, e being Element of S such
that
A1: lc = p^<*e*> by FINSEQ_2:19;
now
let i be set;
assume
A2: i in dom p;
then reconsider i1=i as Element of NAT;
A3: dom p c= dom lc by A1,FINSEQ_1:26;
then consider u being Element of S, a being Element of F such that
A4: lc/.i = a*u by A2,Def10;
take u, a;
thus p/.i = p.i by A2,PARTFUN1:def 6
.= lc.i1 by A1,A2,FINSEQ_1:def 7
.= a*u by A2,A3,A4,PARTFUN1:def 6;
end;
then reconsider p as RightLinearCombination of F by Def10;
A5: len lc = len p +1 by A1,FINSEQ_2:16;
take p;
take e;
thus lc = p^<* e *> by A1;
let i be set such that
A6: i in dom <*e*>;
A7: len lc in dom lc by FINSEQ_5:6;
then
A8: lc/.(len lc) = lc.(len lc) by PARTFUN1:def 6;
dom <*e*> = {1} by FINSEQ_1:2,38;
then
A9: i = 1 by A6,TARSKI:def 1;
consider u being Element of S, a being Element of F such that
A10: lc/.(len lc) = a*u by A7,Def10;
take u, a;
thus <*e*>/.i = <*e*>.i by A6,PARTFUN1:def 6
.= e by A9,FINSEQ_1:40
.= a*u by A1,A5,A10,A8,FINSEQ_1:42;
end;
definition
let R be non empty multLoopStr, A be non empty Subset of R,
L be LinearCombination of A,
E be FinSequence of [:the carrier of R, the carrier of R, the carrier of R:];
pred E represents L means
len E = len L & for i being set st i in
dom L holds L.i = ((E/.i)`1_3)*((E/.i)`2_3)*((E/.i)`3_3) & ((E/.i)`2_3) in A;
end;
theorem
for R being non empty multLoopStr, A being non empty Subset of R, L be
LinearCombination of A ex E be FinSequence of [:the carrier of R, the carrier
of R, the carrier of R:] st E represents L
proof
let R be non empty multLoopStr,A be non empty Subset of R, L be
LinearCombination of A;
set D = [:the carrier of R, the carrier of R, the carrier of R:];
defpred P[set,set] means ex x, y, z being Element of R st $2 = [x, y, z] & y
in A & L/.$1 = x*y*z;
A1: now
let k be Nat;
assume k in Seg len L;
then k in dom L by FINSEQ_1:def 3;
then consider u,v being Element of R, a being Element of A such that
A2: L/.k = u*a*v by Def8;
reconsider b = a as Element of R;
reconsider d = [u, b, v] as Element of D;
take d;
thus P[k, d] by A2;
end;
consider E being FinSequence of D such that
A3: dom E = Seg len L and
A4: for k be Nat st k in Seg len L holds P[k,E/.k] from
RECDEF_1:sch 17(A1);
take E;
thus len E = len L by A3,FINSEQ_1:def 3;
let i being set such that
A5: i in dom L;
reconsider k = i as Element of NAT by A5;
dom L = Seg len L by FINSEQ_1:def 3;
then consider x, y, z being Element of R such that
A6: E/.k = [x, y, z] and
A7: y in A and
A8: L/.k = x*y*z by A4,A5;
thus L.i = ((E/.i)`1_3)*((E/.i)`2_3)*((E/.i)`3_3)
by A5,A6,A8,PARTFUN1:def 6;
thus thesis by A6,A7;
end;
theorem
for R, S being non empty multLoopStr, F being non empty Subset of R,
lc being LinearCombination of F, G being non empty Subset of S, P being
Function of the carrier of R, the carrier of S, E being FinSequence of [:the
carrier of R, the carrier of R, the carrier of R:] st P.:F c= G & E represents
lc holds ex LC being LinearCombination of G st len lc = len LC & for i being
set st i in dom LC
holds LC.i = (P.((E/.i)`1_3))*(P.((E/.i)`2_3))*(P.((E/.i)`3_3))
proof
let R, S be non empty multLoopStr, F be non empty Subset of R, lc be
LinearCombination of F, G be non empty Subset of S, P be Function of the
carrier of R, the carrier of S, E being FinSequence of [:the carrier of R, the
carrier of R, the carrier of R:];
assume
A1: P.:F c= G;
deffunc F(Nat)=(P.((E/.$1)`1_3))*(P.((E/.$1)`2_3))*(P.((E/.$1)`3_3));
consider LC being FinSequence of the carrier of S such that
A2: len LC = len lc and
A3: for k being Nat st k in dom LC holds LC.k = F(k) from FINSEQ_2:sch 1;
assume
A4: E represents lc;
now
let i be set such that
A5: i in dom LC;
dom lc = dom LC by A2,FINSEQ_3:29;
then dom P = the carrier of R & (E/.i)`2_3 in F
by A4,A5,FUNCT_2:def 1;
then P.((E/.i)`2_3) in P.:F by FUNCT_1:def 6;
then reconsider a = P.((E/.i)`2_3) as Element of G by A1;
reconsider u = P.((E/.i)`1_3), v = P.((E/.i)`3_3) as Element of S;
take u, v, a;
LC.i = LC/.i by A5,PARTFUN1:def 6;
hence LC/.i = u*a*v by A3,A5;
end;
then reconsider LC as LinearCombination of G by Def8;
take LC;
thus len lc = len LC by A2;
let i be set;
assume i in dom LC;
hence thesis by A3;
end;
definition
let R be non empty multLoopStr, A be non empty Subset of R,
L be LeftLinearCombination of A,
E be FinSequence of [:the carrier of R, the carrier of R:];
pred E represents L means
len E = len L & for i being set st i in
dom L holds L.i = ((E/.i)`1)*((E/.i)`2) & ((E/.i)`2) in A;
end;
theorem
for R being non empty multLoopStr, A being non empty Subset of R, L be
LeftLinearCombination of A ex E be FinSequence of [:the carrier of R, the
carrier of R:] st E represents L
proof
let R be non empty multLoopStr,A be non empty Subset of R, L be
LeftLinearCombination of A;
set D = [:the carrier of R, the carrier of R:];
defpred P[set,set] means ex x, y being Element of R st $2 = [x, y] & y in A
& L/.$1 = x*y;
A1: now
let k be Nat;
assume k in Seg len L;
then k in dom L by FINSEQ_1:def 3;
then consider u being Element of R, a being Element of A such that
A2: L/.k = u*a by Def9;
reconsider b = a as Element of R;
reconsider d = [u, b] as Element of D;
take d;
thus P[k, d] by A2;
end;
consider E being FinSequence of D such that
A3: dom E = Seg len L and
A4: for k be Nat st k in Seg len L holds P[k,E/.k] from
RECDEF_1:sch 17(A1);
take E;
thus len E = len L by A3,FINSEQ_1:def 3;
let i being set such that
A5: i in dom L;
reconsider k = i as Element of NAT by A5;
dom L = Seg len L by FINSEQ_1:def 3;
then consider x, y being Element of R such that
A6: E/.k = [x, y] and
A7: y in A and
A8: L/.k = x*y by A4,A5;
thus L.i = ((E/.i)`1)*((E/.i)`2) by A5,A6,A8,PARTFUN1:def 6;
thus thesis by A6,A7;
end;
theorem
for R, S being non empty multLoopStr, F being non empty Subset of R,
lc being LeftLinearCombination of F, G being non empty Subset of S, P being
Function of the carrier of R, the carrier of S, E being FinSequence of [:the
carrier of R, the carrier of R:] st P.:F c= G & E represents lc holds ex LC
being LeftLinearCombination of G st len lc = len LC & for i being set st i in
dom LC holds LC.i = (P.(E/.i)`1)*(P.(E/.i)`2)
proof
let R, S be non empty multLoopStr, F be non empty Subset of R, lc be
LeftLinearCombination of F, G be non empty Subset of S, P be Function of the
carrier of R, the carrier of S, E being FinSequence of [:the carrier of R, the
carrier of R:];
assume
A1: P.:F c= G;
deffunc F(Nat)=(P.(E/.$1)`1)*(P.(E/.$1)`2);
consider LC being FinSequence of the carrier of S such that
A2: len LC = len lc and
A3: for k being Nat st k in dom LC holds LC.k = F(k) from FINSEQ_2:sch 1;
assume
A4: E represents lc;
now
let i be set such that
A5: i in dom LC;
dom lc = dom LC by A2,FINSEQ_3:29;
then dom P = the carrier of R & (E/.i)`2 in F by A4,A5,FUNCT_2:def 1;
then P.((E/.i)`2) in P.:F by FUNCT_1:def 6;
then reconsider a = P.(E/.i)`2 as Element of G by A1;
reconsider u = P.(E/.i)`1 as Element of S;
take u, a;
LC.i = LC/.i by A5,PARTFUN1:def 6;
hence LC/.i = u*a by A3,A5;
end;
then reconsider LC as LeftLinearCombination of G by Def9;
take LC;
thus len lc = len LC by A2;
let i be set;
assume i in dom LC;
hence thesis by A3;
end;
definition
let R be non empty multLoopStr, A be non empty Subset of R, L be
RightLinearCombination of A, E be FinSequence of [:the carrier of R, the
carrier of R:];
pred E represents L means
len E = len L & for i being set st i in
dom L holds L.i = ((E/.i)`1)*((E/.i)`2) & ((E/.i)`1) in A;
end;
theorem
for R being non empty multLoopStr, A being non empty Subset of R, L be
RightLinearCombination of A ex E be FinSequence of [:the carrier of R, the
carrier of R:] st E represents L
proof
let R be non empty multLoopStr,A be non empty Subset of R, L be
RightLinearCombination of A;
set D = [:the carrier of R, the carrier of R:];
defpred P[set,set] means ex x, y being Element of R st $2 = [x, y] & x in A
& L/.$1 = x*y;
A1: now
let k be Nat;
assume k in Seg len L;
then k in dom L by FINSEQ_1:def 3;
then consider v being Element of R, a being Element of A such that
A2: L/.k = a*v by Def10;
reconsider b = a as Element of R;
reconsider d = [b, v] as Element of D;
take d;
thus P[k, d] by A2;
end;
consider E being FinSequence of D such that
A3: dom E = Seg len L and
A4: for k be Nat st k in Seg len L holds P[k,E/.k] from
RECDEF_1:sch 17(A1);
take E;
thus len E = len L by A3,FINSEQ_1:def 3;
let i being set such that
A5: i in dom L;
reconsider k = i as Element of NAT by A5;
dom L = Seg len L by FINSEQ_1:def 3;
then consider x, y being Element of R such that
A6: E/.k = [x, y] and
A7: x in A and
A8: L/.k = x*y by A4,A5;
thus L.i = ((E/.i)`1)*((E/.i)`2) by A5,A6,A8,PARTFUN1:def 6;
thus thesis by A6,A7;
end;
theorem
for R, S being non empty multLoopStr, F being non empty Subset of R,
lc being RightLinearCombination of F, G being non empty Subset of S, P being
Function of the carrier of R, the carrier of S, E being FinSequence of [:the
carrier of R, the carrier of R:] st P.:F c= G & E represents lc holds ex LC
being RightLinearCombination of G st len lc = len LC & for i being set st i in
dom LC holds LC.i = (P.(E/.i)`1)*(P.(E/.i)`2)
proof
let R, S be non empty multLoopStr, F be non empty Subset of R, lc be
RightLinearCombination of F, G be non empty Subset of S, P be Function of the
carrier of R, the carrier of S, E being FinSequence of [:the carrier of R, the
carrier of R:];
assume
A1: P.:F c= G;
deffunc F(Nat)=(P.(E/.$1)`1)*(P.(E/.$1)`2);
consider LC being FinSequence of the carrier of S such that
A2: len LC = len lc and
A3: for k being Nat st k in dom LC holds LC.k = F(k) from FINSEQ_2:sch 1;
assume
A4: E represents lc;
now
let i be set such that
A5: i in dom LC;
dom lc = dom LC by A2,FINSEQ_3:29;
then dom P = the carrier of R & (E/.i)`1 in F by A4,A5,FUNCT_2:def 1;
then P.((E/.i)`1) in P.:F by FUNCT_1:def 6;
then reconsider a = P.(E/.i)`1 as Element of G by A1;
reconsider v = P.(E/.i)`2 as Element of S;
take v, a;
LC.i = LC/.i by A5,PARTFUN1:def 6;
hence LC/.i = a*v by A3,A5;
end;
then reconsider LC as RightLinearCombination of G by Def10;
take LC;
thus len lc = len LC by A2;
let i be set;
assume i in dom LC;
hence thesis by A3;
end;
theorem
for R being non empty multLoopStr, A being non empty Subset of R, l
being LinearCombination of A, n being Nat holds l|Seg n is
LinearCombination of A
proof
let R be non empty multLoopStr, A be non empty Subset of R, l be
LinearCombination of A, n being Nat;
reconsider ln = l|Seg n as FinSequence of the carrier of R by FINSEQ_1:18;
now
let i being set such that
A1: i in dom ln;
A2: dom ln c= dom l by RELAT_1:60;
then consider u,v being Element of R, a being Element of A such that
A3: l/.i = u*a*v by A1,Def8;
take u, v, a;
thus ln/.i = ln.i by A1,PARTFUN1:def 6
.= l.i by A1,FUNCT_1:47
.= u*a*v by A1,A2,A3,PARTFUN1:def 6;
end;
hence thesis by Def8;
end;
theorem
for R being non empty multLoopStr, A being non empty Subset of R, l
being LeftLinearCombination of A, n being Nat holds l|Seg n is
LeftLinearCombination of A
proof
let R be non empty multLoopStr, A be non empty Subset of R, l be
LeftLinearCombination of A, n being Nat;
reconsider ln = l|Seg n as FinSequence of the carrier of R by FINSEQ_1:18;
now
let i being set such that
A1: i in dom ln;
A2: dom ln c= dom l by RELAT_1:60;
then consider u being Element of R, a being Element of A such that
A3: l/.i = u*a by A1,Def9;
take u, a;
thus ln/.i = ln.i by A1,PARTFUN1:def 6
.= l.i by A1,FUNCT_1:47
.= u*a by A1,A2,A3,PARTFUN1:def 6;
end;
hence thesis by Def9;
end;
theorem
for R being non empty multLoopStr, A being non empty Subset of R, l
being RightLinearCombination of A, n being Nat holds l|Seg n is
RightLinearCombination of A
proof
let R be non empty multLoopStr, A be non empty Subset of R, l be
RightLinearCombination of A, n being Nat;
reconsider ln = l|Seg n as FinSequence of the carrier of R by FINSEQ_1:18;
now
let i being set such that
A1: i in dom ln;
A2: dom ln c= dom l by RELAT_1:60;
then consider v being Element of R, a being Element of A such that
A3: l/.i = a*v by A1,Def10;
take v, a;
thus ln/.i = ln.i by A1,PARTFUN1:def 6
.= l.i by A1,FUNCT_1:47
.= a*v by A1,A2,A3,PARTFUN1:def 6;
end;
hence thesis by Def10;
end;
begin :: Generated ideals
definition
let L be non empty doubleLoopStr, F be Subset of L;
assume
A1: F is non empty;
func F-Ideal -> Ideal of L means
:Def14:
F c= it & for I being Ideal of L st F c= I holds it c= I;
existence
proof
set Id = { I where I is Subset of L: F c= I & I is Ideal of L };
set I = meet Id;
the carrier of L is Ideal of L by Th10;
then
A2: the carrier of L in Id;
A3: now
let X be set;
assume X in Id;
then ex X9 being Subset of L st X9=X & F c= X9 & X9 is Ideal of L;
hence F c= X;
end;
then F c= I by A2,SETFAM_1:5;
then reconsider I as non empty Subset of L by A1,A2,SETFAM_1:3;
A4: I is add-closed
proof
let x, y being Element of L;
assume
A5: x in I & y in I;
now
let X be set;
assume
A6: X in Id;
then consider X9 being Subset of L such that
A7: X9=X and
F c= X9 and
A8: X9 is Ideal of L;
x in X & y in X by A5,A6,SETFAM_1:def 1;
then x+y in X9 by A7,A8,Def1;
hence {x+y} c= X by A7,ZFMISC_1:31;
end;
then {x+y} c= I by A2,SETFAM_1:5;
hence thesis by ZFMISC_1:31;
end;
A9: I is left-ideal
proof
let p, x being Element of L;
assume
A10: x in I;
now
let X be set;
assume
A11: X in Id;
then consider X9 being Subset of L such that
A12: X9=X and
F c= X9 and
A13: X9 is Ideal of L;
x in X by A10,A11,SETFAM_1:def 1;
then p*x in X9 by A12,A13,Def2;
hence {p*x} c= X by A12,ZFMISC_1:31;
end;
then {p*x} c= I by A2,SETFAM_1:5;
hence thesis by ZFMISC_1:31;
end;
I is right-ideal
proof
let p, x being Element of L;
assume
A14: x in I;
now
let X be set;
assume
A15: X in Id;
then consider X9 being Subset of L such that
A16: X9=X and
F c= X9 and
A17: X9 is Ideal of L;
x in X by A14,A15,SETFAM_1:def 1;
then x*p in X9 by A16,A17,Def3;
hence {x*p} c= X by A16,ZFMISC_1:31;
end;
then {x*p} c= I by A2,SETFAM_1:5;
hence thesis by ZFMISC_1:31;
end;
then reconsider I as Ideal of L by A4,A9;
take I;
now
let X be Ideal of L;
assume F c= X;
then X in Id;
hence I c= X by SETFAM_1:3;
end;
hence thesis by A2,A3,SETFAM_1:5;
end;
uniqueness;
func F-LeftIdeal -> LeftIdeal of L means
:Def15:
F c= it & for I being LeftIdeal of L st F c= I holds it c= I;
existence
proof
set Id = { I where I is Subset of L: F c= I & I is LeftIdeal of L };
set I = meet Id;
the carrier of L is LeftIdeal of L by Th11;
then
A18: the carrier of L in Id;
A19: now
let X be set;
assume X in Id;
then ex X9 being Subset of L st X9=X & F c= X9 & X9 is LeftIdeal of L;
hence F c= X;
end;
then F c= I by A18,SETFAM_1:5;
then reconsider I as non empty Subset of L by A1,A18,SETFAM_1:3;
A20: I is add-closed
proof
let x, y being Element of L;
assume
A21: x in I & y in I;
now
let X be set;
assume
A22: X in Id;
then consider X9 being Subset of L such that
A23: X9=X and
F c= X9 and
A24: X9 is LeftIdeal of L;
x in X & y in X by A21,A22,SETFAM_1:def 1;
then x+y in X9 by A23,A24,Def1;
hence {x+y} c= X by A23,ZFMISC_1:31;
end;
then {x+y} c= I by A18,SETFAM_1:5;
hence thesis by ZFMISC_1:31;
end;
I is left-ideal
proof
let p, x being Element of L;
assume
A25: x in I;
now
let X be set;
assume
A26: X in Id;
then consider X9 being Subset of L such that
A27: X9=X and
F c= X9 and
A28: X9 is LeftIdeal of L;
x in X by A25,A26,SETFAM_1:def 1;
then p*x in X9 by A27,A28,Def2;
hence {p*x} c= X by A27,ZFMISC_1:31;
end;
then {p*x} c= I by A18,SETFAM_1:5;
hence thesis by ZFMISC_1:31;
end;
then reconsider I as LeftIdeal of L by A20;
take I;
now
let X be LeftIdeal of L;
assume F c= X;
then X in Id;
hence I c= X by SETFAM_1:3;
end;
hence thesis by A18,A19,SETFAM_1:5;
end;
uniqueness;
func F-RightIdeal -> RightIdeal of L means
:Def16:
F c= it & for I being RightIdeal of L st F c= I holds it c= I;
existence
proof
set Id = { I where I is Subset of L: F c= I & I is RightIdeal of L };
set I = meet Id;
the carrier of L is RightIdeal of L by Th12;
then
A29: the carrier of L in Id;
A30: now
let X be set;
assume X in Id;
then
ex X9 being Subset of L st X9=X & F c= X9 & X9 is RightIdeal of L;
hence F c= X;
end;
then F c= I by A29,SETFAM_1:5;
then reconsider I as non empty Subset of L by A1,A29,SETFAM_1:3;
A31: I is add-closed
proof
let x, y being Element of L;
assume
A32: x in I & y in I;
now
let X be set;
assume
A33: X in Id;
then consider X9 being Subset of L such that
A34: X9=X and
F c= X9 and
A35: X9 is RightIdeal of L;
x in X & y in X by A32,A33,SETFAM_1:def 1;
then x+y in X9 by A34,A35,Def1;
hence {x+y} c= X by A34,ZFMISC_1:31;
end;
then {x+y} c= I by A29,SETFAM_1:5;
hence thesis by ZFMISC_1:31;
end;
I is right-ideal
proof
let p, x being Element of L;
assume
A36: x in I;
now
let X be set;
assume
A37: X in Id;
then consider X9 being Subset of L such that
A38: X9=X and
F c= X9 and
A39: X9 is RightIdeal of L;
x in X by A36,A37,SETFAM_1:def 1;
then x*p in X9 by A38,A39,Def3;
hence {x*p} c= X by A38,ZFMISC_1:31;
end;
then {x*p} c= I by A29,SETFAM_1:5;
hence thesis by ZFMISC_1:31;
end;
then reconsider I as RightIdeal of L by A31;
take I;
now
let X be RightIdeal of L;
assume F c= X;
then X in Id;
hence I c= X by SETFAM_1:3;
end;
hence thesis by A29,A30,SETFAM_1:5;
end;
uniqueness;
end;
theorem Th44:
for L being non empty doubleLoopStr, I being Ideal of L holds I -Ideal = I
by Def14;
theorem Th45:
for L being non empty doubleLoopStr, I being LeftIdeal of L
holds I-LeftIdeal = I
by Def15;
theorem Th46:
for L being non empty doubleLoopStr, I being RightIdeal of L
holds I-RightIdeal = I
by Def16;
definition
let L be non empty doubleLoopStr, I be Ideal of L;
mode Basis of I -> non empty Subset of L means
it-Ideal = I;
existence
proof
take I;
thus thesis by Th44;
end;
end;
theorem
for L being add-associative right_zeroed right_complementable
distributive non empty doubleLoopStr holds {0.L}-Ideal = {0.L} by Th44;
theorem
for L being left_zeroed right_zeroed add-cancelable distributive non
empty doubleLoopStr holds {0.L}-Ideal = {0.L} by Th44;
theorem
for L being left_zeroed right_zeroed right_add-cancelable
right-distributive non empty doubleLoopStr holds {0.L}-LeftIdeal = {0.L}
by Th45;
theorem
for L being right_zeroed left_add-cancelable left-distributive non
empty doubleLoopStr holds {0.L}-RightIdeal = {0.L} by Th46;
theorem
for L being well-unital non empty doubleLoopStr holds {1.L}-Ideal =
the carrier of L
proof
let L be well-unital non empty doubleLoopStr;
the carrier of L c= {1.L}-Ideal
proof
let x be object;
assume x in the carrier of L;
then reconsider x9=x as Element of L;
1.L in {1.L} & {1.L} c= {1.L}-Ideal by Def14,TARSKI:def 1;
then x9*1.L in {1.L}-Ideal by Def2;
hence thesis;
end;
hence thesis;
end;
theorem
for L being right_unital non empty doubleLoopStr holds {1.L}
-LeftIdeal = the carrier of L
proof
let L be right_unital non empty doubleLoopStr;
the carrier of L c= {1.L}-LeftIdeal
proof
let x be object;
assume x in the carrier of L;
then reconsider x9=x as Element of L;
1.L in {1.L} & {1.L} c= {1.L}-LeftIdeal by Def15,TARSKI:def 1;
then x9*1.L in {1.L}-LeftIdeal by Def2;
hence thesis;
end;
hence thesis;
end;
theorem
for L being left_unital non empty doubleLoopStr holds {1.L}
-RightIdeal = the carrier of L
proof
let L be left_unital non empty doubleLoopStr;
the carrier of L c= {1.L}-RightIdeal
proof
let x be object;
assume x in the carrier of L;
then reconsider x9=x as Element of L;
1.L in {1.L} & {1.L} c= {1.L}-RightIdeal by Def16,TARSKI:def 1;
then (1.L)*x9 in {1.L}-RightIdeal by Def3;
hence thesis;
end;
hence thesis;
end;
theorem
for L being non empty doubleLoopStr holds ([#] L)-Ideal = the carrier of L
by Def14;
theorem
for L being non empty doubleLoopStr holds ([#] L)-LeftIdeal = the
carrier of L
by Def15;
theorem
for L being non empty doubleLoopStr holds ([#] L)-RightIdeal = the
carrier of L
by Def16;
theorem Th57:
for L being non empty doubleLoopStr, A, B being non empty Subset
of L st A c= B holds A-Ideal c= B-Ideal
proof
let L be non empty doubleLoopStr, A,B be non empty Subset of L;
assume
A1: A c= B;
B c= B-Ideal by Def14;
then A c= B-Ideal by A1;
hence thesis by Def14;
end;
theorem
for L being non empty doubleLoopStr, A, B being non empty Subset of L
st A c= B holds A-LeftIdeal c= B-LeftIdeal
proof
let L be non empty doubleLoopStr, A,B be non empty Subset of L;
assume
A1: A c= B;
B c= B-LeftIdeal by Def15;
then A c= B-LeftIdeal by A1;
hence thesis by Def15;
end;
theorem
for L being non empty doubleLoopStr, A, B being non empty Subset of L
st A c= B holds A-RightIdeal c= B-RightIdeal
proof
let L be non empty doubleLoopStr, A,B be non empty Subset of L;
assume
A1: A c= B;
B c= B-RightIdeal by Def16;
then A c= B-RightIdeal by A1;
hence thesis by Def16;
end;
theorem Th60:
for L being add-associative left_zeroed right_zeroed
add-cancelable associative distributive well-unital non empty doubleLoopStr,
F being non empty Subset of L, x being set holds x in F-Ideal iff ex f being
LinearCombination of F st x = Sum f
proof
let L be add-associative right_zeroed add-cancelable left_zeroed associative
distributive well-unital non empty doubleLoopStr, F be non empty Subset of L;
set I = { x where x is Element of L: ex lc being LinearCombination of F st
Sum lc = x };
A1: I c= the carrier of L
proof
let x be object;
assume x in I;
then ex x9 being Element of L st x9=x & ex lc being LinearCombination of F
st Sum lc = x9;
hence thesis;
end;
let x be set;
A2: F c= I
proof
let x be object;
assume
A3: x in F;
then reconsider x as Element of L;
set lc = <* x *>;
now
let i be set;
assume
A4: i in dom lc;
dom lc = {1} by FINSEQ_1:2,38;
then i = 1 by A4,TARSKI:def 1;
then lc.i = x by FINSEQ_1:40
.= 1.L*x
.= 1.L*x*1.L;
hence
ex u,v being Element of L, a being Element of F st lc/.i = u*a*v by A3,A4
,PARTFUN1:def 6;
end;
then reconsider lc as LinearCombination of F by Def8;
Sum lc = x by BINOM:3;
hence thesis;
end;
A5: I c= F-Ideal
proof
defpred P[Nat] means for lc being LinearCombination of F st len lc <= $1
holds Sum lc in F-Ideal;
let x be object;
assume x in I;
then consider x9 being Element of L such that
A6: x9=x and
A7: ex lc being LinearCombination of F st Sum lc = x9;
consider lc being LinearCombination of F such that
A8: Sum lc = x9 by A7;
A9: for k being Nat st P[k] holds P[k+1]
proof
let k be Nat;
assume
A10: P[k];
thus P[k+1]
proof
let lc be LinearCombination of F;
assume
A11: len lc <= k+1;
per cases by A11,NAT_1:8;
suppose
len lc <= k;
hence thesis by A10;
end;
suppose
A12: len lc = k+1;
then lc is non empty;
then consider
q being LinearCombination of F, r being Element of L such
that
A13: lc=q^<*r*> and
A14: <*r*> is LinearCombination of F by Th32;
k+1 = len q + len <*r*> by A12,A13,FINSEQ_1:22
.= len q + 1 by FINSEQ_1:39;
then
A15: Sum q in F-Ideal by A10;
dom <*r*> = {1} by FINSEQ_1:2,38;
then
A16: 1 in dom <*r*> by TARSKI:def 1;
then consider u,v being Element of L, a being Element of F such that
A17: <* r *>/.1 = u*a*v by A14,Def8;
F c= F-Ideal by Def14;
then a in F-Ideal;
then
A18: u*a in F-Ideal by Def2;
A19: <*r*>/.1 = <*r*>.1 by A16,PARTFUN1:def 6;
Sum <* r *> = r by BINOM:3
.= u*a*v by A17,A19,FINSEQ_1:40;
then
A20: Sum <* r *> in F-Ideal by A18,Def3;
Sum lc = Sum q + Sum <* r *> by A13,RLVECT_1:41;
hence thesis by A15,A20,Def1;
end;
end;
end;
A21: P[0]
proof
set y = the Element of F;
let lc be LinearCombination of F;
assume len lc <= 0;
then lc = <*>(the carrier of L);
then
A22: Sum lc = 0.L by RLVECT_1:43;
F c= F-Ideal by Def14;
then
A23: y in F-Ideal;
0.L*y = 0.L by BINOM:1;
hence thesis by A22,A23,Def2;
end;
for k being Nat holds P[k] from NAT_1:sch 2(A21,A9);
then P[len lc];
hence thesis by A6,A8;
end;
reconsider I as non empty Subset of L by A2,A1;
reconsider I9=I as non empty Subset of L;
A24: I9 is add-closed
proof
let x, y be Element of L;
assume that
A25: x in I9 and
A26: y in I9;
consider x9 being Element of L such that
A27: x9=x and
A28: ex lc being LinearCombination of F st Sum lc = x9 by A25;
consider lcx being LinearCombination of F such that
A29: Sum lcx = x9 by A28;
consider y9 being Element of L such that
A30: y9=y and
A31: ex lc being LinearCombination of F st Sum lc = y9 by A26;
consider lcy being LinearCombination of F such that
A32: Sum lcy = y9 by A31;
Sum (lcx^lcy) = x9 + y9 by A29,A32,RLVECT_1:41;
hence thesis by A27,A30;
end;
A33: I9 is right-ideal
proof
let p, x be Element of L;
assume x in I9;
then consider x9 being Element of L such that
A34: x9=x and
A35: ex lc being LinearCombination of F st Sum lc = x9;
consider lcx being LinearCombination of F such that
A36: Sum lcx = x9 by A35;
reconsider lcxp = lcx*p as LinearCombination of F by Th24;
x*p = Sum lcxp by A34,A36,BINOM:5;
hence thesis;
end;
I9 is left-ideal
proof
let p, x be Element of L;
assume x in I9;
then consider x9 being Element of L such that
A37: x9=x and
A38: ex lc being LinearCombination of F st Sum lc = x9;
consider lcx being LinearCombination of F such that
A39: Sum lcx = x9 by A38;
reconsider plcx = p*lcx as LinearCombination of F by Th23;
p*x = Sum plcx by A37,A39,BINOM:4;
hence thesis;
end;
then F-Ideal c= I by A2,A24,A33,Def14;
then
A40: I = F-Ideal by A5;
hereby
assume x in F-Ideal;
then ex x9 being Element of L st x9=x & ex lc being LinearCombination of F
st Sum lc = x9 by A40;
hence ex f being LinearCombination of F st x = Sum f;
end;
assume ex f being LinearCombination of F st x = Sum f;
hence thesis by A40;
end;
theorem Th61:
for L being add-associative left_zeroed right_zeroed
add-cancelable associative distributive well-unital non empty doubleLoopStr,
F being non empty Subset of L, x being set holds x in F-LeftIdeal iff ex f
being LeftLinearCombination of F st x = Sum f
proof
let L be add-associative right_zeroed add-cancelable left_zeroed associative
distributive well-unital non empty doubleLoopStr, F be non empty Subset of L;
set I = { x where x is Element of L: ex lc being LeftLinearCombination of F
st Sum lc = x };
A1: I c= the carrier of L
proof
let x be object;
assume x in I;
then ex x9 being Element of L st x9=x & ex lc being LeftLinearCombination
of F st Sum lc = x9;
hence thesis;
end;
let x be set;
A2: F c= I
proof
let x be object;
assume
A3: x in F;
then reconsider x as Element of L;
set lc = <* x *>;
now
let i be set;
assume
A4: i in dom lc;
dom lc = {1} by FINSEQ_1:2,38;
then i = 1 by A4,TARSKI:def 1;
then lc.i = x by FINSEQ_1:40
.= 1.L*x;
hence
ex u being Element of L, a being Element of F st lc/.i = u*a by A3,A4,
PARTFUN1:def 6;
end;
then reconsider lc as LeftLinearCombination of F by Def9;
Sum lc = x by BINOM:3;
hence thesis;
end;
A5: I c= F-LeftIdeal
proof
defpred P[Nat] means for lc being LeftLinearCombination of F st len lc <=
$1 holds Sum lc in F-LeftIdeal;
let x be object;
assume x in I;
then consider x9 being Element of L such that
A6: x9=x and
A7: ex lc being LeftLinearCombination of F st Sum lc = x9;
consider lc being LeftLinearCombination of F such that
A8: Sum lc = x9 by A7;
A9: for k being Nat st P[k] holds P[k+1]
proof
let k be Nat;
assume
A10: P[k];
thus P[k+1]
proof
let lc be LeftLinearCombination of F;
assume
A11: len lc <= k+1;
per cases by A11,NAT_1:8;
suppose
len lc <= k;
hence thesis by A10;
end;
suppose
A12: len lc = k+1;
then lc is non empty;
then consider
q being LeftLinearCombination of F, r being Element of L
such that
A13: lc=q^<*r*> and
A14: <*r*> is LeftLinearCombination of F by Th33;
k+1 = len q + len <*r*> by A12,A13,FINSEQ_1:22
.= len q + 1 by FINSEQ_1:39;
then
A15: Sum q in F-LeftIdeal by A10;
dom <*r*> = {1} by FINSEQ_1:2,38;
then
A16: 1 in dom <*r*> by TARSKI:def 1;
then consider u being Element of L, a being Element of F such that
A17: <* r *>/.1 = u*a by A14,Def9;
F c= F-LeftIdeal by Def15;
then
A18: a in F-LeftIdeal;
A19: <*r*>/.1 = <*r*>.1 by A16,PARTFUN1:def 6;
Sum <* r *> = r by BINOM:3
.= u*a by A17,A19,FINSEQ_1:40;
then
A20: Sum <* r *> in F-LeftIdeal by A18,Def2;
Sum lc = Sum q + Sum <* r *> by A13,RLVECT_1:41;
hence thesis by A15,A20,Def1;
end;
end;
end;
A21: P[0]
proof
set y = the Element of F;
let lc be LeftLinearCombination of F;
assume len lc <= 0;
then lc = <*>(the carrier of L);
then
A22: Sum lc = 0.L by RLVECT_1:43;
F c= F-LeftIdeal by Def15;
then
A23: y in F-LeftIdeal;
0.L*y = 0.L by BINOM:1;
hence thesis by A22,A23,Def2;
end;
for k being Nat holds P[k] from NAT_1:sch 2(A21,A9);
then P[len lc];
hence thesis by A6,A8;
end;
reconsider I as non empty Subset of L by A2,A1;
reconsider I9=I as non empty Subset of L;
A24: I9 is add-closed
proof
let x, y be Element of L;
assume that
A25: x in I9 and
A26: y in I9;
consider x9 being Element of L such that
A27: x9=x and
A28: ex lc being LeftLinearCombination of F st Sum lc = x9 by A25;
consider lcx being LeftLinearCombination of F such that
A29: Sum lcx = x9 by A28;
consider y9 being Element of L such that
A30: y9=y and
A31: ex lc being LeftLinearCombination of F st Sum lc = y9 by A26;
consider lcy being LeftLinearCombination of F such that
A32: Sum lcy = y9 by A31;
Sum (lcx^lcy) = x9 + y9 by A29,A32,RLVECT_1:41;
hence thesis by A27,A30;
end;
I9 is left-ideal
proof
let p, x be Element of L;
assume x in I9;
then consider x9 being Element of L such that
A33: x9=x and
A34: ex lc being LeftLinearCombination of F st Sum lc = x9;
consider lcx being LeftLinearCombination of F such that
A35: Sum lcx = x9 by A34;
reconsider plcx = p*lcx as LeftLinearCombination of F by Th26;
p*x = Sum plcx by A33,A35,BINOM:4;
hence thesis;
end;
then F-LeftIdeal c= I by A2,A24,Def15;
then
A36: I = F-LeftIdeal by A5;
hereby
assume x in F-LeftIdeal;
then ex x9 being Element of L st x9=x & ex lc being LeftLinearCombination
of F st Sum lc = x9 by A36;
hence ex f being LeftLinearCombination of F st x = Sum f;
end;
assume ex f being LeftLinearCombination of F st x = Sum f;
hence thesis by A36;
end;
theorem Th62:
for L being add-associative left_zeroed right_zeroed
add-cancelable associative distributive well-unital non empty doubleLoopStr,
F being non empty Subset of L, x being set holds x in F-RightIdeal iff ex f
being RightLinearCombination of F st x = Sum f
proof
let L be add-associative left_zeroed right_zeroed add-cancelable associative
distributive well-unital non empty doubleLoopStr, F be non empty Subset of L;
set I = { x where x is Element of L: ex lc being RightLinearCombination of F
st Sum lc = x };
A1: I c= the carrier of L
proof
let x be object;
assume x in I;
then ex x9 being Element of L st x9=x & ex lc being RightLinearCombination
of F st Sum lc = x9;
hence thesis;
end;
let x be set;
A2: F c= I
proof
let x be object;
assume
A3: x in F;
then reconsider x as Element of L;
set lc = <* x *>;
now
let i be set;
assume
A4: i in dom lc;
dom lc = {1} by FINSEQ_1:2,38;
then i = 1 by A4,TARSKI:def 1;
then lc.i = x by FINSEQ_1:40
.= x*1.L;
hence
ex v being Element of L, a being Element of F st lc/.i = a*v by A3,A4,
PARTFUN1:def 6;
end;
then reconsider lc as RightLinearCombination of F by Def10;
Sum lc = x by BINOM:3;
hence thesis;
end;
A5: I c= F-RightIdeal
proof
defpred P[Nat] means for lc being RightLinearCombination of F st len lc <=
$1 holds Sum lc in F-RightIdeal;
let x be object;
assume x in I;
then consider x9 being Element of L such that
A6: x9=x and
A7: ex lc being RightLinearCombination of F st Sum lc = x9;
consider lc being RightLinearCombination of F such that
A8: Sum lc = x9 by A7;
A9: for k being Nat st P[k] holds P[k+1]
proof
let k be Nat;
assume
A10: P[k];
thus P[k+1]
proof
let lc be RightLinearCombination of F;
assume
A11: len lc <= k+1;
per cases by A11,NAT_1:8;
suppose
len lc <= k;
hence thesis by A10;
end;
suppose
A12: len lc = k+1;
then lc is non empty;
then consider
q being RightLinearCombination of F, r being Element of L
such that
A13: lc=q^<*r*> and
A14: <*r*> is RightLinearCombination of F by Th34;
k+1 = len q + len <*r*> by A12,A13,FINSEQ_1:22
.= len q + 1 by FINSEQ_1:39;
then
A15: Sum q in F-RightIdeal by A10;
dom <*r*> = {1} by FINSEQ_1:2,38;
then
A16: 1 in dom <*r*> by TARSKI:def 1;
then consider v being Element of L, a being Element of F such that
A17: <* r *>/.1 = a*v by A14,Def10;
F c= F-RightIdeal by Def16;
then
A18: a in F-RightIdeal;
A19: <*r*>/.1 = <*r*>.1 by A16,PARTFUN1:def 6;
Sum <* r *> = r by BINOM:3
.= a*v by A17,A19,FINSEQ_1:40;
then
A20: Sum <* r *> in F-RightIdeal by A18,Def3;
Sum lc = Sum q + Sum <* r *> by A13,RLVECT_1:41;
hence thesis by A15,A20,Def1;
end;
end;
end;
A21: P[0]
proof
set y = the Element of F;
let lc be RightLinearCombination of F;
assume len lc <= 0;
then lc = <*>(the carrier of L);
then
A22: Sum lc = 0.L by RLVECT_1:43;
F c= F-RightIdeal by Def16;
then
A23: y in F-RightIdeal;
y*0.L = 0.L by BINOM:2;
hence thesis by A22,A23,Def3;
end;
for k being Nat holds P[k] from NAT_1:sch 2(A21,A9);
then P[len lc];
hence thesis by A6,A8;
end;
reconsider I as non empty Subset of L by A2,A1;
reconsider I9=I as non empty Subset of L;
A24: I9 is add-closed
proof
let x, y be Element of L;
assume that
A25: x in I9 and
A26: y in I9;
consider x9 being Element of L such that
A27: x9=x and
A28: ex lc being RightLinearCombination of F st Sum lc = x9 by A25;
consider lcx being RightLinearCombination of F such that
A29: Sum lcx = x9 by A28;
consider y9 being Element of L such that
A30: y9=y and
A31: ex lc being RightLinearCombination of F st Sum lc = y9 by A26;
consider lcy being RightLinearCombination of F such that
A32: Sum lcy = y9 by A31;
Sum (lcx^lcy) = x9 + y9 by A29,A32,RLVECT_1:41;
hence thesis by A27,A30;
end;
I9 is right-ideal
proof
let p, x be Element of L;
assume x in I9;
then consider x9 being Element of L such that
A33: x9=x and
A34: ex lc being RightLinearCombination of F st Sum lc = x9;
consider lcx being RightLinearCombination of F such that
A35: Sum lcx = x9 by A34;
reconsider lcxp = lcx*p as RightLinearCombination of F by Th29;
x*p = Sum lcxp by A33,A35,BINOM:5;
hence thesis;
end;
then F-RightIdeal c= I by A2,A24,Def16;
then
A36: I = F-RightIdeal by A5;
hereby
assume x in F-RightIdeal;
then ex x9 being Element of L st x9=x & ex lc being RightLinearCombination
of F st Sum lc = x9 by A36;
hence ex f being RightLinearCombination of F st x = Sum f;
end;
assume ex f being RightLinearCombination of F st x = Sum f;
hence thesis by A36;
end;
theorem Th63:
for R being add-associative left_zeroed right_zeroed
add-cancelable well-unital associative commutative distributive non empty
doubleLoopStr, F being non empty Subset of R holds F-Ideal = F-LeftIdeal & F
-Ideal = F-RightIdeal
proof
let R be add-associative left_zeroed right_zeroed add-cancelable well-unital
associative commutative distributive non empty doubleLoopStr, F be non empty
Subset of R;
now
let x be object;
hereby
assume x in F-Ideal;
then consider lc being LinearCombination of F such that
A1: x = Sum lc by Th60;
lc is LeftLinearCombination of F by Th31;
hence x in F-LeftIdeal by A1,Th61;
end;
assume x in F-LeftIdeal;
then consider lc being LeftLinearCombination of F such that
A2: x = Sum lc by Th61;
lc is LinearCombination of F by Th25;
hence x in F-Ideal by A2,Th60;
end;
hence F-Ideal = F-LeftIdeal by TARSKI:2;
now
let x be object;
hereby
assume x in F-Ideal;
then consider lc being LinearCombination of F such that
A3: x = Sum lc by Th60;
lc is RightLinearCombination of F by Th31;
hence x in F-RightIdeal by A3,Th62;
end;
assume x in F-RightIdeal;
then consider lc being RightLinearCombination of F such that
A4: x = Sum lc by Th62;
lc is LinearCombination of F by Th28;
hence x in F-Ideal by A4,Th60;
end;
hence thesis by TARSKI:2;
end;
theorem Th64:
for R being add-associative left_zeroed right_zeroed
add-cancelable well-unital associative commutative distributive non empty
doubleLoopStr, a being Element of R holds {a}-Ideal = the set of all
a*r where r is Element
of R
proof
let R be add-associative left_zeroed right_zeroed add-cancelable well-unital
commutative associative distributive non empty doubleLoopStr, a be Element of
R;
set A = {a};
reconsider a9 = a as Element of A by TARSKI:def 1;
set M = the set of all Sum s where s is LinearCombination of A ;
set N = the set of all a*r where r is Element of R ;
A1: for u being object holds u in M implies u in N
proof
let u be object;
assume u in M;
then consider s being LinearCombination of A such that
A2: u = Sum s;
consider f being sequence of the carrier of R such that
A3: Sum s = f.(len s) and
A4: f.0 = 0.R and
A5: for j being Nat,v being Element of R st j < len s & v =
s.(j + 1) holds f.(j + 1) = f.j + v by RLVECT_1:def 12;
defpred P[Element of NAT] means ex r being Element of R st f.($1) = a*r;
A6: now
let j be Element of NAT;
assume that
0 <= j and
A7: j < len s;
thus P[j] implies P[j+1]
proof
assume ex r being Element of R st f.j = a*r;
then consider r1 being Element of R such that
A8: f.j = a*r1;
0 + 1 <= j + 1 & j + 1 <= len s by A7,NAT_1:13;
then j + 1 in Seg(len s) by FINSEQ_1:1;
then
A9: j + 1 in dom s by FINSEQ_1:def 3;
then consider
r2,r3 being Element of R, a9 being Element of A such that
A10: s/.(j+1) = r2*a9*r3 by Def8;
s.(j+1) = s/.(j+1) by A9,PARTFUN1:def 6;
then f.(j+1) = f.j + s/.(j+1) by A5,A7;
then f.(j+1) = a*r1 + r2*a*r3 by A8,A10,TARSKI:def 1
.= a*r1 + a*(r2*r3) by GROUP_1:def 3
.= a*(r1 + r2*r3) by VECTSP_1:def 7;
hence thesis;
end;
end;
f.0 = a*0.R by A4,BINOM:2;
then
A11: P[0];
for k being Element of NAT st 0 <= k & k <= len s holds P[k] from
INT_1:sch 7 (A11,A6);
then ex r being Element of R st Sum s = a*r by A3;
hence thesis by A2;
end;
A12: now
let x be object;
hereby
assume x in {a}-Ideal;
then x in {a}-RightIdeal by Th63;
then consider f being RightLinearCombination of A such that
A13: x = Sum f by Th62;
f is LinearCombination of A by Th28;
hence x in M by A13;
end;
assume x in M;
then ex s being LinearCombination of A st x = Sum s;
hence x in {a}-Ideal by Th60;
end;
for u being object holds u in N implies u in M
proof
let u be object;
assume u in N;
then consider r being Element of R such that
A14: u = a*r;
set s = <* a*r *>;
for i being set st i in dom s ex r,t being Element of R, a being
Element of A st s/.i = r*a*t
proof
let i be set;
A15: len s = 1 by FINSEQ_1:40;
assume i in dom s;
then i in {1} by A15,FINSEQ_1:2,def 3;
then i = 1 by TARSKI:def 1;
then s/.i = a*r by FINSEQ_4:16
.= (r*a9)*1.R;
hence thesis;
end;
then reconsider s as LinearCombination of A by Def8;
Sum s = a*r by BINOM:3;
hence thesis by A14;
end;
then M = N by A1,TARSKI:2;
hence thesis by A12,TARSKI:2;
end;
theorem Th65:
for R being Abelian left_zeroed right_zeroed add-cancelable
well-unital add-associative associative commutative distributive non empty
doubleLoopStr, a,b being Element of R holds {a,b}-Ideal = the set of all
a*r + b*s where r,s
is Element of R
proof
let R be Abelian left_zeroed right_zeroed add-cancelable associative
well-unital add-associative commutative distributive non empty doubleLoopStr,
a,b be Element of R;
set A = {a,b};
reconsider a9=a, b9=b as Element of A by TARSKI:def 2;
set M = the set of all Sum s where s is LinearCombination of A ;
set N = the set of all a*r + b*s where r,s is Element of R ;
A1: for u being object holds u in M implies u in N
proof
let u be object;
assume u in M;
then consider s being LinearCombination of A such that
A2: u = Sum s;
consider f being sequence of the carrier of R such that
A3: Sum s = f.(len s) and
A4: f.0 = 0.R and
A5: for j being Nat,v being Element of R st j < len s & v =
s.(j + 1) holds f.(j + 1) = f.j + v by RLVECT_1:def 12;
defpred P[Element of NAT] means ex r,s being Element of R st f.($1) = a*r
+ b*s;
A6: now
let j be Element of NAT;
assume that
0 <= j and
A7: j < len s;
thus P[j] implies P[j+1]
proof
0 + 1 <= j + 1 & j + 1 <= len s by A7,NAT_1:13;
then j + 1 in Seg(len s) by FINSEQ_1:1;
then
A8: j + 1 in dom s by FINSEQ_1:def 3;
then
A9: s/.(j+1) = s.(j+1) by PARTFUN1:def 6;
assume ex r,s being Element of R st f.j = a*r + b*s;
then consider r1,s1 being Element of R such that
A10: f.j = a*r1 + b*s1;
consider r2,r3 being Element of R, a9 being Element of A such that
A11: s/.(j+1) = r2*a9*r3 by A8,Def8;
per cases by TARSKI:def 2;
suppose
a9 = a;
then f.(j+1) = (a*r1 + b*s1) + r2*a*r3 by A5,A7,A10,A11,A9
.= (a*r1 + a*r2*r3) + b*s1 by RLVECT_1:def 3
.= (a*r1 + a*(r2*r3)) + b*s1 by GROUP_1:def 3
.= a*(r1 + r2*r3) + b*s1 by VECTSP_1:def 7;
hence thesis;
end;
suppose
a9 = b;
then f.(j+1) = (a*r1 + b*s1) + r2*b*r3 by A5,A7,A10,A11,A9
.= a*r1 + (b*s1 + b*r2*r3) by RLVECT_1:def 3
.= a*r1 + (b*s1 + b*(r2*r3)) by GROUP_1:def 3
.= a*r1 + b*(s1 + r2*r3) by VECTSP_1:def 7;
hence thesis;
end;
end;
end;
f.0 = a*0.R by A4,BINOM:2
.= a*0.R + 0.R by RLVECT_1:def 4
.= a*0.R + b*0.R by BINOM:2;
then
A12: P[0];
for k being Element of NAT st 0 <= k & k <= len s holds P[k] from
INT_1:sch 7 (A12,A6);
then ex r,t being Element of R st Sum s = a*r + b*t by A3;
hence thesis by A2;
end;
A13: now
let x be object;
hereby
assume x in {a,b}-Ideal;
then x in {a,b}-RightIdeal by Th63;
then consider f being RightLinearCombination of A such that
A14: x = Sum f by Th62;
f is LinearCombination of A by Th28;
hence x in M by A14;
end;
assume x in M;
then ex s being LinearCombination of A st x = Sum s;
hence x in {a,b}-Ideal by Th60;
end;
for u being object holds u in N implies u in M
proof
let u be object;
assume u in N;
then consider r,t being Element of R such that
A15: u = a*r + b*t;
set s = <* a*r, b*t *>;
for i being set st i in dom s ex r,t being Element of R, a being
Element of A st s/.i = r*a*t
proof
let i be set;
assume
A16: i in dom s;
then i in Seg(len s) by FINSEQ_1:def 3;
then
A17: i in {1,2} by FINSEQ_1:2,44;
per cases by A17,TARSKI:def 2;
suppose
i = 1;
then s/.i = s.1 by A16,PARTFUN1:def 6
.= a*r by FINSEQ_1:44
.= 1.R*a9*r;
hence thesis;
end;
suppose
i = 2;
then s/.i = s.2 by A16,PARTFUN1:def 6
.= b*t by FINSEQ_1:44
.= 1.R*b9*t;
hence thesis;
end;
end;
then reconsider s as LinearCombination of A by Def8;
Sum s = a*r + b*t by Th1;
hence thesis by A15;
end;
then M = N by A1,TARSKI:2;
hence thesis by A13,TARSKI:2;
end;
theorem Th66:
for R being non empty doubleLoopStr, a being Element of R holds
a in {a}-Ideal
proof
let R be non empty doubleLoopStr, a be Element of R;
a in {a} & {a} c= {a}-Ideal by Def14,TARSKI:def 1;
hence thesis;
end;
theorem
for R being Abelian left_zeroed right_zeroed right_complementable
add-associative associative commutative distributive well-unital non empty
doubleLoopStr, A being non empty Subset of R, a being Element of R holds a in
A-Ideal implies {a}-Ideal c= A-Ideal
proof
let R be left_zeroed right_zeroed right_complementable add-associative
associative distributive well-unital commutative Abelian non empty
doubleLoopStr, A be non empty Subset of R, a be Element of R;
assume a in A-Ideal;
then consider s being LinearCombination of A such that
A1: a = Sum s by Th60;
let u be object;
assume u in {a}-Ideal;
then u in the set of all a*r where r is Element of R by Th64;
then consider r being Element of R such that
A2: u = a*r;
set t = s*r;
A3: dom s = dom t by POLYNOM1:def 2;
for i being set st i in dom t ex u,v being Element of R, a being
Element of A st t/.i = u*a*v
proof
let i be set;
assume
A4: i in dom t;
then consider u,v being Element of R, b being Element of A such that
A5: s/.i = u*b*v by A3,Def8;
t/.i = (u*b*v)*r by A3,A4,A5,POLYNOM1:def 2
.= u*b*(v*r) by GROUP_1:def 3;
hence thesis;
end;
then
A6: t is LinearCombination of A by Def8;
Sum t = u by A1,A2,BINOM:5;
hence u in A-Ideal by A6,Th60;
end;
Lm2: for a,b being set holds {a} c= {a,b}
proof
let a,b be set;
let u be object;
assume u in {a};
then u = a by TARSKI:def 1;
hence u in {a,b} by TARSKI:def 2;
end;
theorem
for R being non empty doubleLoopStr, a,b being Element of R holds a in
{a,b}-Ideal & b in {a,b}-Ideal
proof
let R be non empty doubleLoopStr, a,b be Element of R;
{a}-Ideal c= { a,b}-Ideal & a in {a}-Ideal by Lm2,Th57,Th66;
hence a in {a,b}-Ideal;
{b}-Ideal c= {a,b}-Ideal & b in {b}-Ideal by Lm2,Th57,Th66;
hence thesis;
end;
theorem
for R being non empty doubleLoopStr, a,b being Element of R holds {a}
-Ideal c= {a,b}-Ideal & {b}-Ideal c= {a,b}-Ideal by Lm2,Th57;
begin :: Some Operations on Ideals
definition
let R be non empty multMagma, I be Subset of R, a be Element of R;
func a*I -> Subset of R equals
{a*i where i is Element of R : i in I};
coherence
proof
set M = {a*i where i is Element of R : i in I};
M is Subset of R
proof
per cases;
suppose
A1: I is empty;
M is empty
proof
assume M is non empty;
then reconsider M as non empty set;
set b = the Element of M;
b in {a*i where i is Element of R : i in I};
then ex i being Element of R st b = a*i & i in I;
hence thesis by A1;
end;
then for u being object holds u in M implies u in the carrier of R;
hence thesis by TARSKI:def 3;
end;
suppose
I is non empty;
then reconsider I as non empty set;
set j9 = the Element of I;
j9 in I;
then reconsider j = j9 as Element of R;
a*j in {a*i where i is Element of R : i in I};
then reconsider M as non empty set;
for x being object holds x in M implies x in the carrier of R
proof
let x be object;
assume x in M;
then ex i being Element of R st x = a*i & i in I;
hence thesis;
end;
hence thesis by TARSKI:def 3;
end;
end;
hence thesis;
end;
end;
registration
let R be non empty multLoopStr, I be non empty Subset of R,
a be Element of R;
cluster a*I -> non empty;
coherence
proof
set j = the Element of I;
a*j in {a*i where i is Element of R : i in I };
hence thesis;
end;
end;
registration
let R be distributive non empty doubleLoopStr, I be add-closed Subset of R,
a be Element of R;
cluster a*I -> add-closed;
coherence
proof
set M = {a*i where i is Element of R : i in I};
for x,y being Element of R st x in M & y in M holds x+y in M
proof
let x,y be Element of R;
assume that
A1: x in M and
A2: y in M;
consider i being Element of R such that
A3: x = a*i & i in I by A1;
consider j being Element of R such that
A4: y = a*j & j in I by A2;
reconsider k = i + j as Element of R;
k in I & x + y = a*k by A3,A4,Def1,VECTSP_1:def 7;
hence thesis;
end;
hence thesis;
end;
end;
registration
let R be associative non empty doubleLoopStr, I be right-ideal Subset of R,
a be Element of R;
cluster a*I -> right-ideal;
coherence
proof
set M = {a*i where i is Element of R : i in I};
for y,x being Element of R st x in M holds x*y in M
proof
let y,x be Element of R;
assume x in M;
then consider i being Element of R such that
A1: x = a*i & i in I;
x*y = a*(i*y) & i*y in I by A1,Def3,GROUP_1:def 3;
hence thesis;
end;
hence thesis;
end;
end;
theorem Th70:
for R being right_zeroed left_add-cancelable left-distributive
non empty doubleLoopStr, I being non empty Subset of R holds 0.R*I = {0.R}
proof
let R be right_zeroed left_add-cancelable left-distributive non empty
doubleLoopStr, I be non empty Subset of R;
A1: now
set j = the Element of I;
let u be object;
assume u in {0.R};
then
A2: u = 0.R by TARSKI:def 1;
0.R*j = 0.R by BINOM:1;
hence u in 0.R*I by A2;
end;
now
let u be object;
assume u in 0.R*I;
then ex i being Element of R st u = 0.R*i & i in I;
then u = 0.R by BINOM:1;
hence u in {0.R} by TARSKI:def 1;
end;
hence thesis by A1,TARSKI:2;
end;
theorem
for R being left_unital non empty doubleLoopStr, I being Subset of R
holds 1.R*I = I
proof
let R be left_unital non empty doubleLoopStr, I be Subset of R;
A1: now
let u be object;
assume
A2: u in I;
then reconsider u9 = u as Element of R;
1.R*u9 = u9;
hence u in 1.R*I by A2;
end;
now
let u be object;
assume u in 1.R*I;
then ex i being Element of R st u = 1.R*i & i in I;
hence u in I;
end;
hence thesis by A1,TARSKI:2;
end;
definition
let R be addMagma, I,J be Subset of R;
func I + J -> Subset of R equals
{a + b where a,b is Element of R : a in I & b in J};
coherence
proof
set M = {a + b where a,b is Element of R : a in I & b in J};
M is Subset of R
proof
per cases;
suppose
A1: I is empty or J is empty;
now
per cases by A1;
case
A2: I is empty;
M is empty
proof
assume M is non empty;
then reconsider M as non empty set;
set x = the Element of M;
x in {a + b where a,b is Element of R : a in I & b in J};
then ex a,b being Element of R st x = a + b & a in I & b in J;
hence thesis by A2;
end;
then for u being object
holds u in M implies u in the carrier of R;
hence thesis by TARSKI:def 3;
end;
case
A3: J is empty;
M is empty
proof
assume M is non empty;
then reconsider M as non empty set;
set x = the Element of M;
x in {a + b where a,b is Element of R : a in I & b in J};
then ex a,b being Element of R st x = a + b & a in I & b in J;
hence thesis by A3;
end;
then for u being object
holds u in M implies u in the carrier of R;
hence thesis by TARSKI:def 3;
end;
end;
hence thesis;
end;
suppose
A4: I is non empty & J is non empty;
then reconsider J as non empty set;
reconsider I as non empty set by A4;
set x9 = the Element of I;
set y9 = the Element of J;
x9 in I & y9 in J;
then reconsider x = x9,y = y9 as Element of R;
x+y in {a + b where a,b is Element of R : a in I & b in J};
then reconsider M as non empty set;
for x being object holds x in M implies x in the carrier of R
proof
let x be object;
assume x in M;
then ex a,b being Element of R st x = a + b & a in I & b in J;
hence thesis;
end;
hence thesis by TARSKI:def 3;
end;
end;
hence thesis;
end;
end;
registration
let R be non empty addLoopStr, I,J be non empty Subset of R;
cluster I + J -> non empty;
coherence
proof
{x + y where x,y is Element of R : x in I & y in J} is non empty
proof
set y = the Element of J;
set x = the Element of I;
x+y in {a + b where a,b is Element of R : a in I & b in J};
hence thesis;
end;
hence thesis;
end;
end;
definition
let R be Abelian non empty addLoopStr, I,J be Subset of R;
redefine func I + J;
commutativity
proof
now
let I,J be Subset of R;
A1: now
let u be object;
assume u in J+I;
then ex a,b being Element of R st u = a + b & a in J & b in I;
hence u in I+J;
end;
now
let u be object;
assume u in I+J;
then ex a,b being Element of R st u = a + b & a in I & b in J;
hence u in J+I;
end;
hence I + J = J + I by A1,TARSKI:2;
end;
hence thesis;
end;
end;
registration
let R be Abelian add-associative non empty addLoopStr, I,J be add-closed
Subset of R;
cluster I + J -> add-closed;
coherence
proof
set M = {a + b where a,b is Element of R : a in I & b in J};
for x,y being Element of R st x in M & y in M holds x+y in M
proof
let x,y be Element of R;
assume that
A1: x in M and
A2: y in M;
consider a9,b9 being Element of R such that
A3: x = a9 + b9 and
A4: a9 in I & b9 in J by A1;
consider c,d being Element of R such that
A5: y = c + d and
A6: c in I & d in J by A2;
A7: (a9 + c) + (b9 + d) = ((a9 + c) + b9) + d by RLVECT_1:def 3
.= (c + x) + d by A3,RLVECT_1:def 3
.= x + y by A5,RLVECT_1:def 3;
a9 + c in I & b9 + d in J by A4,A6,Def1;
hence thesis by A7;
end;
hence thesis;
end;
end;
registration
let R be left-distributive non empty doubleLoopStr, I,J be right-ideal
Subset of R;
cluster I + J -> right-ideal;
coherence
proof
set M = {a + b where a,b is Element of R : a in I & b in J};
for y,x being Element of R st x in M holds x*y in M
proof
let y,x be Element of R;
assume x in M;
then consider a9,b9 being Element of R such that
A1: x = a9 + b9 and
A2: a9 in I & b9 in J;
A3: (a9*y) + (b9*y) = x*y by A1,VECTSP_1:def 3;
a9*y in I & b9*y in J by A2,Def3;
hence thesis by A3;
end;
hence thesis;
end;
end;
registration
let R be right-distributive non empty doubleLoopStr, I,J be left-ideal
Subset of R;
cluster I + J -> left-ideal;
coherence
proof
set M = {a + b where a,b is Element of R : a in I & b in J};
for y,x being Element of R st x in M holds y*x in M
proof
let y,x be Element of R;
assume x in M;
then consider a9,b9 being Element of R such that
A1: x = a9 + b9 and
A2: a9 in I & b9 in J;
A3: (y*a9) + (y*b9) = y*x by A1,VECTSP_1:def 2;
y*a9 in I & y*b9 in J by A2,Def2;
hence thesis by A3;
end;
hence thesis;
end;
end;
theorem
for R being add-associative non empty addLoopStr, I,J,K being Subset
of R holds I + (J + K) = (I + J) + K
proof
let R be add-associative non empty addLoopStr, I,J,K be Subset of R;
A1: now
let u be object;
assume u in (I + J) + K;
then consider a,b being Element of R such that
A2: u = a + b and
A3: a in I+J and
A4: b in K;
consider c,d being Element of R such that
A5: a = c + d and
A6: c in I and
A7: d in J by A3;
d + b in {a9 + b9 where a9,b9 is Element of R : a9 in J & b9 in K} by A4,A7
;
then
c+(d + b) in {a9 + b9 where a9,b9 is Element of R : a9 in I & b9 in J
+K} by A6;
hence u in I + (J + K) by A2,A5,RLVECT_1:def 3;
end;
now
let u be object;
assume u in I + (J + K);
then consider a,b being Element of R such that
A8: u = a + b and
A9: a in I and
A10: b in J+K;
consider c,d being Element of R such that
A11: b = c + d and
A12: c in J and
A13: d in K by A10;
a + c in {a9 + b9 where a9,b9 is Element of R : a9 in I & b9 in J} by A9
,A12;
then (a+ c) + d in {a9 + b9 where a9,b9 is Element of R : a9 in I+J & b9
in K}by A13;
hence u in (I + J) + K by A8,A11,RLVECT_1:def 3;
end;
hence thesis by A1,TARSKI:2;
end;
theorem Th73:
for R being left_zeroed right_zeroed right_add-cancelable
right-distributive non empty doubleLoopStr, I,J being right-ideal non empty
Subset of R holds I c= I + J
proof
let R be left_zeroed right_add-cancelable right_zeroed right-distributive
non empty doubleLoopStr, I,J be right-ideal non empty Subset of R;
let u be object;
assume u in I;
then reconsider u9 = u as Element of I;
0.R is Element of J by Th3;
then u9 + 0.R in {a + b where a,b is Element of R : a in I & b in J};
hence u in I + J by RLVECT_1:def 4;
end;
theorem Th74:
for R being left_zeroed right_add-cancelable right-distributive
non empty doubleLoopStr, I,J being right-ideal non empty Subset of R holds
J c= I + J
proof
let R be left_zeroed right_add-cancelable right-distributive non empty
doubleLoopStr, I,J be right-ideal non empty Subset of R;
let u be object;
assume u in J;
then reconsider u9 = u as Element of J;
0.R is Element of I by Th3;
then 0.R + u9 in {a + b where a,b is Element of R : a in I & b in J};
hence u in I + J by ALGSTR_1:def 2;
end;
theorem
for R being non empty addLoopStr, I,J being Subset of R, K being
add-closed Subset of R st I c= K & J c= K holds I + J c= K
proof
let R be non empty addLoopStr, I,J be Subset of R, K being add-closed (
Subset of R);
assume
A1: I c= K & J c= K;
let u be object;
assume u in I + J;
then ex i,j being Element of R st u = i + j & i in I & j in J;
hence u in K by A1,Def1;
end;
theorem
for R being Abelian left_zeroed right_zeroed add-cancelable
well-unital add-associative associative commutative distributive non empty
doubleLoopStr, a,b being Element of R holds {a,b}-Ideal = {a}-Ideal + {b}
-Ideal
proof
let R be Abelian left_zeroed right_zeroed add-cancelable well-unital
add-associative associative commutative distributive non empty doubleLoopStr,
a,b be Element of R;
A1: now
let u be object;
assume u in {a,b}-Ideal;
then u in the set of all a*r + b*s where r,s is Element of R by Th65;
then consider r,s being Element of R such that
A2: u = a*r + b*s;
b*s in the set of all b*v where v is Element of R ;
then reconsider b9 = b*s as Element of {b}-Ideal by Th64;
a*r in the set of all a*v where v is Element of R ;
then reconsider a9 = a*r as Element of {a}-Ideal by Th64;
a9 + b9 in {x + y where x,y is Element of R : x in {a}-Ideal & y in {b
}-Ideal};
hence u in {a}-Ideal + {b}-Ideal by A2;
end;
now
let u be object;
assume u in {a}-Ideal + {b}-Ideal;
then consider x,y being Element of R such that
A3: u = x + y and
A4: x in {a}-Ideal and
A5: y in {b}-Ideal;
y in the set of all b*v where v is Element of R by A5,Th64;
then
A6: ex s being Element of R st y = b*s;
x in the set of all a*v where v is Element of R by A4,Th64;
then ex r being Element of R st x = a*r;
then
u in the set of all a*v + b*d where v,d is Element of R by A3,A6;
hence u in {a,b}-Ideal by Th65;
end;
hence thesis by A1,TARSKI:2;
end;
definition
let R be non empty 1-sorted, I,J be Subset of R;
redefine func I /\ J -> Subset of R equals
{ x where x is Element of R : x in I & x in J };
coherence
proof
I /\ J is Subset of R;
hence thesis;
end;
compatibility
proof
defpred Q[set] means $1 in J;
defpred P[set] means $1 in I;
set X = { x where x is Element of R : P[x] & Q[x] }, Y = { x where x is
Element of R : P[x] }, Z = { x where x is Element of R : Q[x] };
A1: Y = I by DOMAIN_1:22;
X = Y /\ Z from DOMAIN_1:sch 10;
hence thesis by A1,DOMAIN_1:22;
end;
end;
registration
let R be right_zeroed left_add-cancelable left-distributive non empty
doubleLoopStr, I,J be left-ideal non empty Subset of R;
cluster I /\ J -> non empty;
coherence
proof
0.R in I & 0.R in J by Th2;
then 0.R in { x where x is Element of R : x in I & x in J };
hence thesis;
end;
end;
registration
let R be non empty addLoopStr, I,J be add-closed Subset of R;
cluster I /\ J -> add-closed for Subset of R;
coherence
proof
set M = { x where x is Element of R : x in I & x in J };
M = I /\ J;
then reconsider M as Subset of R;
for x,y being Element of R st x in M & y in M holds x+y in M
proof
let x,y be Element of R;
assume that
A1: x in M and
A2: y in M;
consider c being Element of R such that
A3: c = y and
A4: c in I & c in J by A2;
consider a being Element of R such that
A5: x = a and
A6: a in I & a in J by A1;
a + c in I & a + c in J by A6,A4,Def1;
hence thesis by A5,A3;
end;
hence thesis;
end;
end;
registration
let R be non empty multLoopStr, I,J be left-ideal Subset of R;
cluster I /\ J -> left-ideal for Subset of R;
coherence
proof
set M = { x where x is Element of R : x in I & x in J };
M = I /\ J;
then reconsider M as Subset of R;
for y,x being Element of R st x in M holds y*x in M
proof
let y,x be Element of R;
assume x in M;
then consider a being Element of R such that
A1: x = a and
A2: a in I & a in J;
y*a in I & y*a in J by A2,Def2;
hence thesis by A1;
end;
hence thesis;
end;
end;
registration
let R be non empty multLoopStr, I,J be right-ideal Subset of R;
cluster I /\ J -> right-ideal for Subset of R;
coherence
proof
set M = { x where x is Element of R : x in I & x in J };
M = I /\ J;
then reconsider M as Subset of R;
for y,x being Element of R st x in M holds x*y in M
proof
let y,x be Element of R;
assume x in M;
then consider a being Element of R such that
A1: x = a and
A2: a in I & a in J;
a*y in I & a*y in J by A2,Def3;
hence thesis by A1;
end;
hence thesis;
end;
end;
theorem
for R being Abelian left_zeroed right_zeroed right_complementable
left_unital add-associative left-distributive non empty doubleLoopStr, I
being add-closed left-ideal non empty Subset of R, J being Subset of R, K
being non empty Subset of R st J c= I holds I /\ (J + K) = J + (I /\ K)
proof
let R be Abelian left_zeroed right_zeroed right_complementable left_unital
add-associative left-distributive non empty doubleLoopStr, I be add-closed
left-ideal non empty Subset of R, J be Subset of R, K be non empty Subset of
R;
assume
A1: J c= I;
A2: now
let u be object;
assume u in J + (I /\ K);
then consider j,ik being Element of R such that
A3: u = j + ik and
A4: j in J and
A5: ik in (I /\ K);
A6: ex z being Element of R st z = ik & z in I & z in K by A5;
then reconsider k9 = ik as Element of K;
u = j + k9 by A3;
then
A7: u in J + K by A4;
reconsider j9 = j as Element of I by A1,A4;
reconsider i9 = ik as Element of I by A6;
u = j9 + i9 by A3;
then u in I by Def1;
hence u in I /\ (J + K) by A7;
end;
now
let u be object;
assume u in I /\ (J + K);
then consider v being Element of R such that
A8: u = v and
A9: v in I and
A10: v in J + K;
consider j,k being Element of R such that
A11: v = j + k and
A12: j in J and
A13: k in K by A10;
reconsider j9 = j as Element of I by A1,A12;
-j9 in I by Th13;
then (j9 + k) + -j9 in I by A9,A11,Def1;
then (j9 + -j9) + k in I by RLVECT_1:def 3;
then 0.R + k in I by RLVECT_1:5;
then k in I by ALGSTR_1:def 2;
then k in (I /\ K) by A13;
hence u in J + (I /\ K) by A8,A11,A12;
end;
hence thesis by A2,TARSKI:2;
end;
definition
let R be non empty doubleLoopStr, I,J be Subset of R;
func I *' J -> Subset of R equals
{ Sum s where s is FinSequence of the
carrier of R : for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in I & b in J};
coherence
proof
set M = {Sum s where s is FinSequence of the carrier of R : for i being
Element of NAT st 1 <= i & i <= len s ex a,b being Element of R st s.i = a*b &
a in I & b in J};
now
let u be object;
assume u in M;
then
ex s being FinSequence of the carrier of R st u = Sum s & for i being
Element of NAT st 1 <= i & i <= len s ex a,b being Element of R st s.i = a*b &
a in I & b in J;
hence u in the carrier of R;
end;
then reconsider M as Subset of R by TARSKI:def 3;
M is Subset of R;
hence thesis;
end;
end;
registration
let R be non empty doubleLoopStr, I,J be Subset of R;
cluster I *' J -> non empty;
coherence
proof
set M = {Sum s where s is FinSequence of the carrier of R : for i being
Element of NAT st 1 <= i & i <= len s ex a,b being Element of R st s.i = a*b &
a in I & b in J};
M is non empty
proof
set p = <*>(the carrier of R);
for i being Element of NAT st 1 <= i & i <= len p ex a,b being
Element of R st p.i = a*b & a in I & b in J;
then Sum p in M;
hence thesis;
end;
hence thesis;
end;
end;
definition
let R be commutative non empty doubleLoopStr, I,J be Subset of R;
redefine func I *' J;
commutativity
proof
now
let I,J be Subset of R;
A1: now
let u be object;
assume u in J*'I;
then consider s being FinSequence of the carrier of R such that
A2: u = Sum s and
A3: for i being Element of NAT st 1 <= i & i <= len s ex a,b
being Element of R st s.i = a*b & a in J & b in I;
for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in I & b in J
proof
let i be Element of NAT;
assume 1 <= i & i <= len s;
then
ex a,b being Element of R st s.i = a*b & a in J & b in I by A3;
hence thesis;
end;
hence u in I*'J by A2;
end;
now
let u be object;
assume u in I*'J;
then consider s being FinSequence of the carrier of R such that
A4: u = Sum s and
A5: for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in I & b in J;
for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in J & b in I
proof
let i be Element of NAT;
assume 1 <= i & i <= len s;
then ex a,b being Element of R st s.i = a*b & a in I & b in J by A5;
hence thesis;
end;
hence u in J*'I by A4;
end;
hence I *' J = J *' I by A1,TARSKI:2;
end;
hence thesis;
end;
end;
registration
let R be right_zeroed add-associative non empty doubleLoopStr, I,J be
Subset of R;
cluster I *' J -> add-closed;
coherence
proof
set M = {Sum s where s is FinSequence of the carrier of R : for i being
Element of NAT st 1 <= i & i <= len s ex a,b being Element of R st s.i = a*b &
a in I & b in J};
M = I *' J;
then reconsider M as non empty Subset of R;
for x,y being Element of R st x in M & y in M holds x+y in M
proof
let x,y be Element of R;
assume that
A1: x in M and
A2: y in M;
consider s being FinSequence of the carrier of R such that
A3: x = Sum s and
A4: for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in I & b in J by A1;
consider t being FinSequence of the carrier of R such that
A5: y = Sum t and
A6: for i being Element of NAT st 1 <= i & i <= len t ex a,b being
Element of R st t.i = a*b & a in I & b in J by A2;
set q = s^t;
A7: now
let i be Element of NAT;
assume that
A8: 1 <= i and
A9: i <= len q;
thus ex a,r being Element of R st q.i = a*r & a in I & r in J
proof
per cases;
suppose
A10: i <= len s;
then i in Seg(len s) by A8,FINSEQ_1:1;
then i in dom s by FINSEQ_1:def 3;
then q.i = s.i by FINSEQ_1:def 7;
hence thesis by A4,A8,A10;
end;
suppose
A11: len s < i;
then reconsider j = i - len s as Element of NAT by INT_1:5;
len s - len s < j by A11,XREAL_1:9;
then
A12: 1 <= j by NAT_1:14;
i <= len s + len t by A9,FINSEQ_1:22;
then
A13: j <= (len s + len t) - len s by XREAL_1:9;
t.j = q.i by A9,A11,FINSEQ_1:24;
hence thesis by A6,A12,A13;
end;
end;
end;
Sum q = x + y by A3,A5,RLVECT_1:41;
hence thesis by A7;
end;
hence thesis;
end;
end;
registration
let R be right_zeroed left_add-cancelable associative left-distributive non
empty doubleLoopStr, I,J be right-ideal Subset of R;
cluster I *' J -> right-ideal;
coherence
proof
set M = {Sum s where s is FinSequence of the carrier of R : for i being
Element of NAT st 1 <= i & i <= len s ex a,b being Element of R st s.i = a*b &
a in I & b in J};
M = I *' J;
then reconsider M as non empty Subset of R;
for y,x being Element of R st x in M holds x*y in M
proof
let y,x be Element of R;
assume x in M;
then consider s being FinSequence of the carrier of R such that
A1: x = Sum s and
A2: for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in I & b in J;
set q = s*y;
A3: Seg(len q) = dom q by FINSEQ_1:def 3
.= dom s by POLYNOM1:def 2
.= Seg(len s) by FINSEQ_1:def 3;
then
A4: len q = len s by FINSEQ_1:6;
A5: now
let i be Element of NAT;
assume
A6: 1 <= i & i <= len q;
then consider c,r9 being Element of R such that
A7: s.i = c*r9 and
A8: c in I & r9 in J by A2,A4;
i in Seg(len s) by A3,A6,FINSEQ_1:1;
then
A9: i in dom s by FINSEQ_1:def 3;
then
A10: s/.i = c*r9 by A7,PARTFUN1:def 6;
i in Seg(len q) by A6,FINSEQ_1:1;
then i in dom q by FINSEQ_1:def 3;
then
A11: q.i = q/.i by PARTFUN1:def 6
.= (c*r9)*y by A9,A10,POLYNOM1:def 2
.= c*(r9*y) by GROUP_1:def 3;
thus ex b,r being Element of R st q.i = b*r & b in I & r in J
proof
take c,r9*y;
thus thesis by A8,A11,Def3;
end;
end;
Sum q = Sum s*y by BINOM:5;
hence thesis by A1,A5;
end;
hence thesis;
end;
end;
registration
let R be left_zeroed right_add-cancelable associative right-distributive
non empty doubleLoopStr, I,J be left-ideal Subset of R;
cluster I *' J -> left-ideal;
coherence
proof
set M = {Sum s where s is FinSequence of the carrier of R : for i being
Element of NAT st 1 <= i & i <= len s ex a,b being Element of R st s.i = a*b &
a in I & b in J};
M = I *' J;
then reconsider M as non empty Subset of R;
for y,x being Element of R st x in M holds y*x in M
proof
let y,x be Element of R;
assume x in M;
then consider s being FinSequence of the carrier of R such that
A1: x = Sum s and
A2: for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in I & b in J;
set q = y*s;
A3: Seg(len q) = dom q by FINSEQ_1:def 3
.= dom s by POLYNOM1:def 1
.= Seg(len s) by FINSEQ_1:def 3;
then
A4: len q = len s by FINSEQ_1:6;
A5: now
let i be Element of NAT;
assume
A6: 1 <= i & i <= len q;
then consider c,r9 being Element of R such that
A7: s.i = c*r9 and
A8: c in I & r9 in J by A2,A4;
i in Seg(len s) by A3,A6,FINSEQ_1:1;
then
A9: i in dom s by FINSEQ_1:def 3;
then
A10: s/.i = c*r9 by A7,PARTFUN1:def 6;
i in Seg(len q) by A6,FINSEQ_1:1;
then i in dom q by FINSEQ_1:def 3;
then
A11: q.i = q/.i by PARTFUN1:def 6
.= y*(c*r9) by A9,A10,POLYNOM1:def 1
.= (y*c)*r9 by GROUP_1:def 3;
thus ex b,r being Element of R st q.i = b*r & b in I & r in J
proof
take y*c,r9;
thus thesis by A8,A11,Def2;
end;
end;
Sum q = y*Sum s by BINOM:4;
hence thesis by A1,A5;
end;
hence thesis;
end;
end;
theorem
for R being left_zeroed right_zeroed left_add-cancelable
left-distributive non empty doubleLoopStr, I being non empty Subset of R
holds {0.R} *' I = {0.R}
proof
let R be left_zeroed right_zeroed left_add-cancelable left-distributive non
empty doubleLoopStr, I be non empty Subset of R;
A1: now
let u be object;
assume u in ({0.R}) *' I;
then consider s being FinSequence of the carrier of R such that
A2: Sum s = u and
A3: for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in {0.R} & b in I;
now
per cases;
case
len s = 0;
then s = <*>(the carrier of R);
hence Sum s = 0.R by RLVECT_1:43;
end;
case
len s <> 0;
then 1 <= len s by NAT_1:14;
then 1 in Seg(len s) by FINSEQ_1:1;
then
A4: 1 in dom s by FINSEQ_1:def 3;
A5: for i being Element of NAT st i in dom s holds s/.i = 0.R
proof
let i be Element of NAT;
assume
A6: i in dom s;
then i in Seg(len s) by FINSEQ_1:def 3;
then 1 <= i & i <= len s by FINSEQ_1:1;
then consider a,b being Element of R such that
A7: s.i = a*b and
A8: a in {0.R} and
b in I by A3;
A9: a = 0.R by A8,TARSKI:def 1;
s/.i = a*b by A6,A7,PARTFUN1:def 6;
hence thesis by A9,BINOM:1;
end;
then
for i being Element of NAT st i in dom s & i <> 1 holds s/.i = 0. R;
hence Sum s = s/.1 by A4,POLYNOM2:3
.= 0.R by A4,A5;
end;
end;
hence u in {0.R} by A2,TARSKI:def 1;
end;
now
reconsider o = 0.R as Element of {0.R} by TARSKI:def 1;
set a = the Element of I;
let u be object;
assume
A10: u in {0.R};
set q = <* 0.R*a *>;
A11: len q = 1 & q.1 = 0.R*a by FINSEQ_1:40;
A12: for i being Element of NAT st 1 <= i & i <= len q holds ex b,r being
Element of R st q.i = b*r & b in {0.R} & r in I
proof
let i be Element of NAT;
assume 1 <= i & i <= len q;
then q.i = o*a by A11,XXREAL_0:1;
hence thesis;
end;
Sum q = 0.R*a by BINOM:3
.= 0.R by BINOM:1
.= u by A10,TARSKI:def 1;
hence u in {0.R} *' I by A12;
end;
hence thesis by A1,TARSKI:2;
end;
theorem Th79:
for R being left_zeroed right_zeroed add-cancelable distributive
non empty doubleLoopStr,
I being add-closed right-ideal non empty Subset of R,
J being add-closed left-ideal non empty Subset of R holds I *' J c= I /\ J
proof
let R be left_zeroed right_zeroed add-cancelable distributive non empty
doubleLoopStr, I be add-closed right-ideal non empty Subset of R, J be
add-closed left-ideal non empty Subset of R;
let u be object;
assume u in I *' J;
then consider s being FinSequence of the carrier of R such that
A1: Sum s = u and
A2: for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in I & b in J;
consider f being sequence of the carrier of R such that
A3: Sum s = f.(len s) and
A4: f.0 = 0.R and
A5: for j being Nat, v being Element of R st j < len s & v
= s.(j + 1) holds f.(j + 1) = f.j + v by RLVECT_1:def 12;
defpred P[Element of NAT] means f.$1 in I & f.$1 in J;
A6: now
let j be Element of NAT;
assume that
0 <= j and
A7: j < len s;
thus P[j] implies P[j+1]
proof
A8: j + 1 <= len s & 0 + 1 <= j + 1 by A7,NAT_1:13;
then j + 1 in Seg(len s) by FINSEQ_1:1;
then j + 1 in dom s by FINSEQ_1:def 3;
then
A9: s.(j+1) = s/.(j+1) by PARTFUN1:def 6;
ex a,b being Element of R st s.(j+1) = a*b & a in I & b in J by A2,A8;
then
A10: s/.(j+1) in I & s/.(j+1) in J by A9,Def2,Def3;
assume
A11: f.j in I & f.j in J;
f.(j+1) = f.j + s/.(j+1) by A5,A7,A9;
hence thesis by A11,A10,Def1;
end;
end;
A12: P[0] by A4,Th2,Th3;
for j being Element of NAT st 0 <= j & j <= len s holds P[j] from
INT_1:sch 7(A12,A6);
then Sum s in I & Sum s in J by A3;
hence u in I /\ J by A1;
end;
theorem Th80:
for R being Abelian left_zeroed right_zeroed add-cancelable
add-associative associative distributive non empty doubleLoopStr, I,J,K being
right-ideal non empty Subset of R holds I *' (J + K) = (I *' J) + (I *' K)
proof
let R be Abelian left_zeroed right_zeroed add-cancelable add-associative
associative distributive non empty doubleLoopStr, I,J,K be right-ideal non
empty Subset of R;
A1: now
let u be object;
assume u in I *' (J + K);
then consider s being FinSequence of the carrier of R such that
A2: Sum s = u and
A3: for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in I & b in J+K;
consider f being sequence of the carrier of R such that
A4: Sum s = f.(len s) and
A5: f.0 = 0.R and
A6: for j being Nat,v being Element of R st j < len s & v =
s .(j + 1) holds f.(j + 1) = f.j + v by RLVECT_1:def 12;
defpred P[Element of NAT] means ex x,y being Element of R st f.$1=x+y & x
in I*'J & y in I*'K;
A7: now
let n be Element of NAT;
assume that
0 <= n and
A8: n < len s;
thus P[n] implies P[n+1]
proof
assume ex x,y being Element of R st f.n = x+y & x in I*'J & y in I*' K;
then consider x,y being Element of R such that
A9: f.n = x+y and
A10: x in I*'J and
A11: y in I*'K;
consider p being FinSequence of the carrier of R such that
A12: Sum p = y and
A13: for i being Element of NAT st 1 <= i & i <= len p ex a,b
being Element of R st p.i = a*b & a in I & b in K by A11;
consider q being FinSequence of the carrier of R such that
A14: Sum q = x and
A15: for i being Element of NAT st 1 <= i & i <= len q ex a,b
being Element of R st q.i = a*b & a in I & b in J by A10;
A16: 0 + 1 <= n + 1 & n + 1 <= len s by A8,NAT_1:13;
then consider a,b being Element of R such that
A17: s.(n+1) = a*b and
A18: a in I and
A19: b in J+K by A3;
consider c,d being Element of R such that
A20: b = c + d and
A21: c in J and
A22: d in K by A19;
set q1 = q^<*a*c*>, p1 = p^<*a*d*>;
n + 1 in Seg(len s) by A16,FINSEQ_1:1;
then
A23: n + 1 in dom s by FINSEQ_1:def 3;
then
A24: s.(n+1) = s/.(n+1) by PARTFUN1:def 6;
A25: len p1 = len p + len(<*a*d*>) by FINSEQ_1:22
.= len p + 1 by FINSEQ_1:40;
for i being Element of NAT st 1 <= i & i <= len p1 ex a,b being
Element of R st p1.i = a*b & a in I & b in K
proof
let i be Element of NAT;
assume that
A26: 1 <= i and
A27: i <= len p1;
per cases;
suppose
i = len p1;
hence thesis by A18,A22,A25,FINSEQ_1:42;
end;
suppose
i <> len p1;
then i < len p1 by A27,XXREAL_0:1;
then
A28: i <= len p by A25,NAT_1:13;
then consider a,b being Element of R such that
A29: p.i = a*b and
A30: a in I & b in K by A13,A26;
i in Seg(len p) by A26,A28,FINSEQ_1:1;
then i in dom p by FINSEQ_1:def 3;
then p1.i = a*b by A29,FINSEQ_1:def 7;
hence thesis by A30;
end;
end;
then
A31: Sum p1 in I *' K;
A32: len q1 = len q + len(<*a*c*>) by FINSEQ_1:22
.= len q + 1 by FINSEQ_1:40;
for i being Element of NAT st 1 <= i & i <= len q1 ex a,b being
Element of R st q1.i = a*b & a in I & b in J
proof
let i be Element of NAT;
assume that
A33: 1 <= i and
A34: i <= len q1;
per cases;
suppose
i = len q1;
hence thesis by A18,A21,A32,FINSEQ_1:42;
end;
suppose
i <> len q1;
then i < len q1 by A34,XXREAL_0:1;
then
A35: i <= len q by A32,NAT_1:13;
then consider a,b being Element of R such that
A36: q.i = a*b and
A37: a in I & b in J by A15,A33;
i in Seg(len q) by A33,A35,FINSEQ_1:1;
then i in dom q by FINSEQ_1:def 3;
then q1.i = a*b by A36,FINSEQ_1:def 7;
hence thesis by A37;
end;
end;
then
A38: Sum q1 in {Sum t where t is FinSequence of the carrier of R: for
i being Element of NAT st 1 <= i & i <= len t ex a,b being Element of R st t.i
= a*b & a in I & b in J};
A39: s/.(n+1) = a*(c + d) by A23,A17,A20,PARTFUN1:def 6
.= a*c + a*d by VECTSP_1:def 7;
Sum q1 + Sum p1 = (Sum q + Sum <*a*c*>) + Sum p1 by RLVECT_1:41
.= (Sum q + a*c) + Sum p1 by BINOM:3
.= (Sum q + a*c) + (Sum p + Sum <*a*d*>) by RLVECT_1:41
.= (Sum q + a*c) + (Sum p + a*d) by BINOM:3
.= ((Sum q + a*c) + Sum p) + a*d by RLVECT_1:def 3
.= (a*c + (Sum q + Sum p)) + a*d by RLVECT_1:def 3
.= f.n + (a*c + a*d) by A9,A14,A12,RLVECT_1:def 3
.= f.(n+1) by A6,A8,A24,A39;
hence thesis by A38,A31;
end;
end;
A40: P[0]
proof
take 0.R,0.R;
thus thesis by A5,Th3,RLVECT_1:def 4;
end;
for n being Element of NAT st 0 <= n &n <= len s holds P[n] from
INT_1:sch 7(A40,A7);
then ex x,y being Element of R st Sum s = x+y & x in I*'J & y in I*'K by A4
;
hence u in (I *' J) + (I *' K) by A2;
end;
now
let u be object;
assume u in (I *' J) + (I *' K);
then consider a,b being Element of R such that
A41: u = a + b and
A42: a in I*'J and
A43: b in I*'K;
consider p being FinSequence of the carrier of R such that
A44: b = Sum p and
A45: for i being Element of NAT st 1 <= i & i <= len p ex a,b being
Element of R st p.i = a*b & a in I & b in K by A43;
consider q being FinSequence of the carrier of R such that
A46: a = Sum q and
A47: for i being Element of NAT st 1 <= i & i <= len q ex a,b being
Element of R st q.i = a*b & a in I & b in J by A42;
set s = p^q;
A48: for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in I & b in J+K
proof
let i be Element of NAT;
assume that
A49: 1 <= i and
A50: i <= len s;
i in Seg(len s) by A49,A50,FINSEQ_1:1;
then
A51: i in dom s by FINSEQ_1:def 3;
now
per cases;
case
A52: i <= len p;
then consider a,b being Element of R such that
A53: p.i = a*b and
A54: a in I and
A55: b in K by A45,A49;
i in Seg(len p) by A49,A52,FINSEQ_1:1;
then i in dom p by FINSEQ_1:def 3;
then
A56: s.i=a*b by A53,FINSEQ_1:def 7
.=a*(0.R + b) by ALGSTR_1:def 2;
0.R in J by Th3;
then
0.R + b in {a9+b9 where a9,b9 is Element of R:a9 in J & b9 in K
} by A55;
hence thesis by A54,A56;
end;
case
i > len p;
then not i in Seg(len p) by FINSEQ_1:1;
then not i in dom p by FINSEQ_1:def 3;
then consider n being Nat such that
A57: n in dom q and
A58: i = len p + n by A51,FINSEQ_1:25;
n in Seg(len q) by A57,FINSEQ_1:def 3;
then 1 <= n & n <= len q by FINSEQ_1:1;
then consider a,b being Element of R such that
A59: q.n = a*b and
A60: a in I and
A61: b in J by A47,A57;
0.R in K by Th3;
then
A62: b + 0.R in {a9+b9 where a9,b9 is Element of R:a9 in J & b9 in K
} by A61;
s.i = q.n by A57,A58,FINSEQ_1:def 7
.= a*(b + 0.R) by A59,RLVECT_1:def 4;
hence thesis by A60,A62;
end;
end;
hence thesis;
end;
Sum s = u by A41,A46,A44,RLVECT_1:41;
hence u in I *' (J + K) by A48;
end;
hence thesis by A1,TARSKI:2;
end;
theorem Th81:
for R being Abelian left_zeroed right_zeroed add-cancelable
add-associative commutative associative distributive non empty doubleLoopStr,
I,J being right-ideal non empty Subset of R holds (I + J) *' (I /\ J) c= I *' J
proof
let R be Abelian left_zeroed right_zeroed add-cancelable add-associative
commutative associative distributive non empty doubleLoopStr, I,J be
right-ideal non empty Subset of R;
A1: now
let u be object;
assume u in (I *' (I /\ J)) + (J *' (I /\ J));
then consider a,b being Element of R such that
A2: u = a + b and
A3: a in (I *' (I /\ J)) and
A4: b in (J *' (I /\ J));
consider s being FinSequence of the carrier of R such that
A5: b = Sum s and
A6: for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in J & b in (I /\ J) by A4;
for i being Element of NAT st 1 <= i & i <= len s ex x,y being
Element of R st s.i = x*y & x in I & y in J
proof
let i be Element of NAT;
assume 1 <= i & i <= len s;
then
A7: ex x,y being Element of R st s.i = x*y & x in J & y in (I /\ J) by A6;
I /\ J c= I by XBOOLE_1:17;
hence thesis by A7;
end;
then
A8: Sum s in {Sum t where t is FinSequence of the carrier of R : for i
being Element of NAT st 1 <= i & i <= len t ex a,b being Element of R st t.i =
a*b & a in I & b in J};
consider q being FinSequence of the carrier of R such that
A9: a = Sum q and
A10: for i being Element of NAT st 1 <= i & i <= len q ex a,b being
Element of R st q.i = a*b & a in I & b in (I /\ J) by A3;
for i being Element of NAT st 1 <= i & i <= len q ex x,y being Element
of R st q.i = x*y & x in I & y in J
proof
let i be Element of NAT;
assume 1 <= i & i <= len q;
then
A11: ex x,y being Element of R st q.i = x*y & x in I & y in (I /\ J) by A10;
I /\ J c= J by XBOOLE_1:17;
hence thesis by A11;
end;
then Sum q in {Sum t where t is FinSequence of the carrier of R : for i
being Element of NAT st 1 <= i & i <= len t ex a,b being Element of R st t.i =
a*b & a in I & b in J};
hence u in I *' J by A2,A9,A5,A8,Def1;
end;
(I + J) *' (I /\ J) = (I *' (I /\ J)) + (J *' (I /\ J)) by Th80;
hence thesis by A1;
end;
theorem
for R being right_zeroed left_add-cancelable left-distributive non
empty doubleLoopStr, I,J being add-closed left-ideal non empty Subset of R
holds (I + J) *' (I /\ J) c= I /\ J
proof
let R be right_zeroed left_add-cancelable left-distributive non empty
doubleLoopStr, I,J be add-closed left-ideal non empty Subset of R;
let u be object;
assume u in (I + J) *' (I /\ J);
then consider s being FinSequence of the carrier of R such that
A1: u = Sum s and
A2: for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in I+J & b in I/\J;
consider f being sequence of the carrier of R such that
A3: Sum s = f.(len s) and
A4: f.0 = 0.R and
A5: for j being Nat,v being Element of R st j < len s & v =
s .(j + 1) holds f.(j + 1) = f.j + v by RLVECT_1:def 12;
defpred P[Element of NAT] means f.$1 in I/\J;
A6: now
let n be Element of NAT;
assume that
0 <= n and
A7: n < len s;
thus P[n] implies P[n+1]
proof
A8: 0 + 1 <= n + 1 & n + 1 <= len s by A7,NAT_1:13;
then n + 1 in Seg(len s) by FINSEQ_1:1;
then n + 1 in dom s by FINSEQ_1:def 3;
then
A9: s.(n+1) = s/.(n+1) by PARTFUN1:def 6;
assume
A10: f.n in I /\ J;
ex x,y being Element of R st s.(n+1) = x*y & x in I+J & y in I/\J
by A2,A8;
then s/.(n+1) in I /\ J by A9,Def2;
then f.n + s/.(n+1) in I /\ J by A10,Def1;
hence thesis by A5,A7,A9;
end;
end;
A11: P[0] by A4,Th2;
for n being Element of NAT st 0 <= n &n <= len s holds P[n] from
INT_1:sch 7 (A11,A6);
hence u in I /\ J by A1,A3;
end;
definition
let R be addMagma, I,J be Subset of R;
pred I,J are_co-prime means
I + J = the carrier of R;
end;
theorem Th83:
for R being left_zeroed left_unital non empty doubleLoopStr,
I,J being non empty Subset of R st I,J are_co-prime holds
I /\ J c= (I + J) *' (I /\ J)
proof
let R be left_zeroed left_unital non empty doubleLoopStr, I,J be non empty
Subset of R;
assume I,J are_co-prime;
then
A1: I + J = the carrier of R;
let u be object;
assume
A2: u in I /\ J;
then reconsider u9 = u as Element of R;
set q = <*1.R*u9*>;
A3: len q = 1 by FINSEQ_1:39;
A4: for i being Element of NAT st 1 <= i & i <= len q ex x,y being Element
of R st q.i = x*y & x in I+J & y in I/\J
proof
let i be Element of NAT;
assume
A5: 1 <= i & i <= len q;
take 1.R,u9;
i = 1 by A3,A5,XXREAL_0:1;
hence thesis by A1,A2,FINSEQ_1:40;
end;
Sum q = 1.R*u9 by BINOM:3
.= u9;
hence u in (I + J) *' (I /\ J) by A4;
end;
theorem
for R being Abelian left_zeroed right_zeroed add-cancelable
add-associative left_unital commutative associative distributive non empty
doubleLoopStr, I being add-closed left-ideal right-ideal non empty Subset of
R, J being add-closed left-ideal non empty Subset of R st I,J are_co-prime
holds I *' J = I /\ J
proof
let R be left_zeroed right_zeroed add-cancelable add-associative Abelian
commutative associative distributive left_unital non empty doubleLoopStr, I
be add-closed left-ideal right-ideal non empty Subset of R, J be add-closed
left-ideal non empty Subset of R;
A1: I *' J c= I /\ J by Th79;
assume I,J are_co-prime;
then
A2: I /\ J c= (I + J) *' (I /\ J) by Th83;
(I + J) *' (I /\ J) c= I *' J by Th81;
then I /\ J c= I *' J by A2;
hence thesis by A1;
end;
definition
let R be non empty multMagma, I,J be Subset of R;
func I % J -> Subset of R equals
{a where a is Element of R: a*J c= I};
coherence
proof
set M = {a where a is Element of R: a*J c= I};
for x being object holds x in M implies x in the carrier of R
proof
let x be object;
assume x in M;
then ex a being Element of R st x = a & a*J c= I;
hence thesis;
end;
hence thesis by TARSKI:def 3;
end;
end;
registration
let R be right_zeroed left_add-cancelable left-distributive non empty
doubleLoopStr, I,J be left-ideal non empty Subset of R;
cluster I % J -> non empty;
coherence
proof
set M = {a where a is Element of R: a*J c= I};
0.R in I by Th2;
then for u being object holds u in {0.R} implies u in I by TARSKI:def 1;
then
A1: {0.R} c= I;
0.R*J = {0.R} by Th70;
then 0.R in M by A1;
hence thesis;
end;
end;
registration
let R be right_zeroed left_add-cancelable left-distributive non empty
doubleLoopStr, I,J be add-closed left-ideal non empty Subset of R;
cluster I % J -> add-closed;
coherence
proof
set M = {a where a is Element of R: a*J c= I};
M = I % J;
then reconsider M as non empty Subset of R;
for x,y being Element of R st x in M & y in M holds x+y in M
proof
let x,y be Element of R;
assume that
A1: x in M and
A2: y in M;
consider b being Element of R such that
A3: y = b and
A4: b*J c= I by A2;
consider a being Element of R such that
A5: x = a and
A6: a*J c= I by A1;
now
let u be object;
assume u in (a+b)*J;
then consider c being Element of R such that
A7: u = (a + b)*c and
A8: c in J;
A9: b*c in {b*i where i is Element of R : i in J} by A8;
u = a*c + b*c & a*c in a*J by A7,A8,VECTSP_1:def 3;
hence u in I by A6,A4,A9,Def1;
end;
then (a+b)*J c= I;
hence thesis by A5,A3;
end;
hence thesis;
end;
end;
registration
let R be right_zeroed left_add-cancelable left-distributive associative
commutative non empty doubleLoopStr, I,J be left-ideal non empty Subset of R;
cluster I % J -> left-ideal;
coherence
proof
set M = {a where a is Element of R: a*J c= I};
M = I % J;
then reconsider M as non empty Subset of R;
for y,x being Element of R st x in M holds y*x in M
proof
let y,x be Element of R;
assume x in M;
then consider a being Element of R such that
A1: x = a and
A2: a*J c= I;
now
let u be object;
assume u in (y*a)*J;
then consider c being Element of R such that
A3: u = (y*a)*c and
A4: c in J;
y*c in J by A4,Def2;
then
A5: a*(y*c) in {a*i where i is Element of R : i in J};
u = a*(y*c) by A3,GROUP_1:def 3;
hence u in I by A2,A5;
end;
then (y*a)*J c= I;
hence thesis by A1;
end;
hence thesis;
end;
cluster I % J -> right-ideal;
coherence;
end;
theorem
for R being non empty multLoopStr, I being right-ideal non empty
Subset of R, J being Subset of R holds I c= I % J
proof
let R be non empty multLoopStr, I be right-ideal non empty Subset of R,
J be Subset of R;
let u be object;
assume
A1: u in I;
then reconsider u9 = u as Element of R;
now
let v be object;
assume v in u9*J;
then ex j being Element of R st v = u9*j & j in J;
hence v in I by A1,Def3;
end;
then u9*J c= I;
hence u in (I % J);
end;
theorem
for R being right_zeroed left_add-cancelable left-distributive non
empty doubleLoopStr, I being add-closed left-ideal non empty Subset of R, J
being Subset of R holds (I % J) *' J c= I
proof
let R be right_zeroed left_add-cancelable left-distributive non empty
doubleLoopStr, I be add-closed left-ideal non empty Subset of R, J be Subset
of R;
let u be object;
assume u in (I % J) *' J;
then consider s being FinSequence of the carrier of R such that
A1: Sum s = u and
A2: for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in I%J & b in J;
consider f being sequence of the carrier of R such that
A3: Sum s = f.(len s) and
A4: f.0 = 0.R and
A5: for j being Nat, v being Element of R st j < len s & v
= s.(j + 1) holds f.(j + 1) = f.j + v by RLVECT_1:def 12;
defpred P[Element of NAT] means f.$1 in I;
A6: now
let j be Element of NAT;
assume that
0 <= j and
A7: j < len s;
thus P[j] implies P[j+1]
proof
A8: j + 1 <= len s & 0 + 1 <= j + 1 by A7,NAT_1:13;
then consider a,b being Element of R such that
A9: s.(j+1) = a*b and
A10: a in I%J and
A11: b in J by A2;
j + 1 in Seg(len s) by A8,FINSEQ_1:1;
then j + 1 in dom s by FINSEQ_1:def 3;
then
A12: s.(j+1) = s/.(j+1) by PARTFUN1:def 6;
then
A13: f.(j+1) = f.j + s/.(j+1) by A5,A7;
assume
A14: f.j in I;
consider d being Element of R such that
A15: a = d and
A16: d*J c= I by A10;
a*b in {d*i where i is Element of R : i in J} by A11,A15;
hence thesis by A14,A12,A13,A9,A16,Def1;
end;
end;
A17: P[0] by A4,Th2;
for j being Element of NAT st 0 <= j & j <= len s holds P[j] from
INT_1:sch 7(A17,A6);
hence u in I by A1,A3;
end;
theorem Th87:
for R being left_zeroed right_add-cancelable right-distributive
non empty doubleLoopStr, I being add-closed right-ideal non empty Subset of
R, J being Subset of R holds (I % J) *' J c= I
proof
let R be left_zeroed right_add-cancelable right-distributive non empty
doubleLoopStr, I be add-closed right-ideal non empty Subset of R, J be
Subset of R;
let u be object;
assume u in (I % J) *' J;
then consider s being FinSequence of the carrier of R such that
A1: Sum s = u and
A2: for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in I%J & b in J;
consider f being sequence of the carrier of R such that
A3: Sum s = f.(len s) and
A4: f.0 = 0.R and
A5: for j being Nat, v being Element of R st j < len s & v
= s.(j + 1) holds f.(j + 1) = f.j + v by RLVECT_1:def 12;
defpred P[Element of NAT] means f.$1 in I;
A6: now
let j be Element of NAT;
assume that
0 <= j and
A7: j < len s;
thus P[j] implies P[j+1]
proof
A8: j + 1 <= len s & 0 + 1 <= j + 1 by A7,NAT_1:13;
then consider a,b being Element of R such that
A9: s.(j+1) = a*b and
A10: a in I%J and
A11: b in J by A2;
j + 1 in Seg(len s) by A8,FINSEQ_1:1;
then j + 1 in dom s by FINSEQ_1:def 3;
then
A12: s.(j+1) = s/.(j+1) by PARTFUN1:def 6;
then
A13: f.(j+1) = f.j + s/.(j+1) by A5,A7;
assume
A14: f.j in I;
consider d being Element of R such that
A15: a = d and
A16: d*J c= I by A10;
a*b in {d*i where i is Element of R : i in J} by A11,A15;
hence thesis by A14,A12,A13,A9,A16,Def1;
end;
end;
A17: P[0] by A4,Th3;
for j being Element of NAT st 0 <= j & j <= len s holds P[j] from
INT_1:sch 7 (A17,A6);
hence u in I by A1,A3;
end;
theorem
for R being left_zeroed right_add-cancelable right-distributive
commutative associative non empty doubleLoopStr, I being add-closed
right-ideal non empty Subset of R, J,K being Subset of R holds (I % J) % K =
I % (J *' K)
proof
let R be left_zeroed right_add-cancelable right-distributive commutative
associative non empty doubleLoopStr, I be add-closed right-ideal non empty
Subset of R, J,K be Subset of R;
A1: now
let u be object;
assume u in (I % J) % K;
then consider a being Element of R such that
A2: u = a and
A3: a*K c= I % J;
now
let v be object;
assume v in a*(J *' K);
then consider b being Element of R such that
A4: v = a*b and
A5: b in J *' K;
consider s being FinSequence of the carrier of R such that
A6: Sum s = b and
A7: for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in J & b in K by A5;
set q = a*s;
A8: dom q = dom s by POLYNOM1:def 1;
A9: Seg(len q) = dom q by FINSEQ_1:def 3
.= dom s by POLYNOM1:def 1
.= Seg(len s) by FINSEQ_1:def 3;
then
A10: len q = len s by FINSEQ_1:6;
for j being Element of NAT st 1 <= j & j <= len q holds ex c,d
being Element of R st q.j = c*d & c in I%J & d in J
proof
let j be Element of NAT;
assume
A11: 1 <= j & j <= len q;
then consider c,d being Element of R such that
A12: s.j = c*d and
A13: c in J and
A14: d in K by A7,A10;
A15: a*d in {a*b9 where b9 is Element of R : b9 in K} by A14;
j in Seg(len s) by A9,A11,FINSEQ_1:1;
then
A16: j in dom s by FINSEQ_1:def 3;
then
A17: s/.j = c*d by A12,PARTFUN1:def 6;
q.j = q/.j by A8,A16,PARTFUN1:def 6
.= a*(c*d) by A16,A17,POLYNOM1:def 1
.= (a*d)*c by GROUP_1:def 3;
hence thesis by A3,A13,A15;
end;
then
A18: Sum q in {Sum t where t is FinSequence of the carrier of R : for i
being Element of NAT st 1 <= i & i <= len t ex a,b being Element of R st t.i =
a*b & a in I%J & b in J};
A19: (I % J) *' J c= I by Th87;
Sum q = v by A4,A6,BINOM:4;
hence v in I by A18,A19;
end;
then a*(J*'K) c= I;
hence u in I % (J *' K) by A2;
end;
now
let u be object;
assume u in I % (J *' K);
then consider a being Element of R such that
A20: u = a and
A21: a*(J *' K) c= I;
now
let v be object;
assume v in a*K;
then consider b being Element of R such that
A22: v = a*b and
A23: b in K;
now
let z be object;
assume z in (a*b)*J;
then consider c being Element of R such that
A24: z = (a*b)*c and
A25: c in J;
A26: z = a*(c*b) by A24,GROUP_1:def 3;
set q = <*c*b*>;
A27: len q = 1 by FINSEQ_1:40;
A28: for i being Element of NAT st 1 <= i & i <= len q ex x,y being
Element of R st q.i = x*y & x in J & y in K
proof
let i be Element of NAT;
assume 1 <= i & i <= len q;
then q.i = q.1 by A27,XXREAL_0:1
.= c*b by FINSEQ_1:40;
hence thesis by A23,A25;
end;
Sum q = c*b by BINOM:3;
then c*b in {Sum t where t is FinSequence of the carrier of R : for i
being Element of NAT st 1 <= i & i <= len t ex a,b being Element of R st t.i =
a*b & a in J & b in K} by A28;
then z in {a*f where f is Element of R : f in J*'K} by A26;
hence z in I by A21;
end;
then (a*b)*J c= I;
hence v in I % J by A22;
end;
then a*K c= I % J;
hence u in (I % J) % K by A20;
end;
hence thesis by A1,TARSKI:2;
end;
theorem
for R being non empty multLoopStr, I,J,K being Subset of R holds (J /\
K) % I = (J % I) /\ (K % I)
proof
let R be non empty multLoopStr, I,J,K be Subset of R;
A1: now
let u be object;
assume u in (J /\ K) % I;
then consider a being Element of R such that
A2: u = a and
A3: a*I c= (J /\ K);
now
let v be object;
assume v in a*I;
then v in J /\ K by A3;
then ex x being Element of R st v = x & x in J & x in K;
hence v in K;
end;
then a*I c= K;
then
A4: u in (K % I) by A2;
now
let v be object;
assume v in a*I;
then v in J /\ K by A3;
then ex x being Element of R st v = x & x in J & x in K;
hence v in J;
end;
then a*I c= J;
then u in (J % I) by A2;
hence u in (J % I) /\ (K % I) by A4;
end;
now
let u be object;
assume u in (J % I) /\ (K % I);
then
A5: ex x being Element of R st x = u & x in (J % I) & x in (K % I);
then consider a being Element of R such that
A6: u = a and
A7: a*I c= J;
ex b being Element of R st u = b & b*I c= K by A5;
then for v be object st v in a*I holds v in J /\ K by A6,A7;
then a*I c= J /\ K;
hence u in (J /\ K) % I by A6;
end;
hence thesis by A1,TARSKI:2;
end;
theorem
for R being left_zeroed right_zeroed right_add-cancelable
right-distributive non empty doubleLoopStr, I being add-closed Subset of R, J
,K being right-ideal non empty Subset of R holds I % (J + K) = (I % J) /\ (I
% K)
proof
let R be left_zeroed right_zeroed right_add-cancelable right-distributive
non empty doubleLoopStr, I be add-closed Subset of R, J,K be right-ideal non
empty Subset of R;
A1: now
let u be object;
assume u in I % (J + K);
then consider a being Element of R such that
A2: u = a and
A3: a*(J+K) c= I;
now
let u be object;
assume u in a*J;
then
A4: ex j being Element of R st u = a*j & j in J;
J c= J+K by Th73;
then u in {a*j9 where j9 is Element of R : j9 in J+K} by A4;
hence u in I by A3;
end;
then a*J c= I;
then
A5: u in (I % J) by A2;
now
let u be object;
assume u in a*K;
then
A6: ex j being Element of R st u = a*j & j in K;
K c= J+K by Th74;
then u in {a*j9 where j9 is Element of R : j9 in J+K} by A6;
hence u in I by A3;
end;
then a*K c= I;
then u in (I % K) by A2;
hence u in (I % J) /\ (I % K) by A5;
end;
now
let u be object;
assume u in (I % J) /\ (I % K);
then
A7: ex x being Element of R st u = x & x in (I % J) & x in (I % K);
then consider a being Element of R such that
A8: u = a and
A9: a*J c= I;
consider b being Element of R such that
A10: u = b and
A11: b*K c= I by A7;
now
let v be object;
assume v in a*(J+K);
then consider j being Element of R such that
A12: v = a*j and
A13: j in J+K;
consider x9,y being Element of R such that
A14: j = x9 + y and
A15: x9 in J & y in K by A13;
A16: a*x9 in a*J & b*y in {b*j9 where j9 is Element of R : j9 in K} by A15;
v = a*x9 + b*y by A8,A10,A12,A14,VECTSP_1:def 2;
hence v in I by A9,A11,A16,Def1;
end;
then a*(J+K) c= I;
hence u in I % (J + K) by A8;
end;
hence thesis by A1,TARSKI:2;
end;
definition
let R be well-unital non empty doubleLoopStr, I be Subset of R;
func sqrt I -> Subset of R equals
{a where a is Element of R: ex n being
Element of NAT st a|^n in I};
coherence
proof
set M = {a where a is Element of R:ex n being Element of NAT st a|^n in I};
for x being object holds x in M implies x in the carrier of R
proof
let x be object;
assume x in M;
then
ex a being Element of R st a = x & ex n being Element of NAT st a|^n
in I;
hence thesis;
end;
hence thesis by TARSKI:def 3;
end;
end;
registration
let R be well-unital non empty doubleLoopStr, I be non empty Subset of R;
cluster sqrt I -> non empty;
coherence
proof
set M ={a where a is Element of R: ex n being Element of NAT st a|^n in I};
M is non empty
proof
set a = the Element of I;
a|^1 = a by BINOM:8;
then a in M;
hence thesis;
end;
hence thesis;
end;
end;
registration
let R be Abelian add-associative left_zeroed right_zeroed commutative
associative add-cancelable distributive well-unital non empty doubleLoopStr,
I be add-closed right-ideal non empty Subset of R;
cluster sqrt I -> add-closed;
coherence
proof
set M ={a where a is Element of R: ex n being Element of NAT st a|^n in I};
M = sqrt I;
then reconsider M as non empty Subset of R;
for x,y being Element of R st x in M & y in M holds x+y in M
proof
let x,y be Element of R;
assume that
A1: x in M and
A2: y in M;
consider a being Element of R such that
A3: x = a and
A4: ex n being Element of NAT st a|^n in I by A1;
consider n being Element of NAT such that
A5: a|^n in I by A4;
consider b being Element of R such that
A6: y = b and
A7: ex m being Element of NAT st b|^m in I by A2;
consider m being Element of NAT such that
A8: b|^m in I by A7;
set p = ((a,b) In_Power (n+m));
consider f being sequence of the carrier of R such that
A9: Sum p = f.(len p) and
A10: f.0 = 0.R and
A11: for j being Nat, v being Element of R st j < len p &
v = p.(j + 1) holds f.(j + 1) = f.j + v by RLVECT_1:def 12;
defpred P[Element of NAT] means f.$1 in I;
A12: for i being Element of NAT st 1 <= i & i <= len p holds p.i in I
proof
let i be Element of NAT;
assume that
A13: 1 <= i and
A14: i <= len p;
set r = i - 1;
set l = (n+m) - r;
1 - 1 <= i - 1 by A13,XREAL_1:9;
then reconsider r as Element of NAT by INT_1:3;
i <= (n+m) + 1 by A14,BINOM:def 7;
then r <= ((n+m) + 1) - 1 by XREAL_1:9;
then r - r <= (n+m) - r by XREAL_1:9;
then reconsider l as Element of NAT by INT_1:3;
i in Seg(len p) by A13,A14,FINSEQ_1:1;
then
A15: i in dom p by FINSEQ_1:def 3;
then
A16: p.i = p/.i by PARTFUN1:def 6
.= ((n+m) choose r)*a|^l*b|^r by A15,BINOM:def 7;
per cases;
suppose
n <= l;
then consider k being Nat such that
A17: l = n + k by NAT_1:10;
reconsider k as Element of NAT by ORDINAL1:def 12;
a|^l = (a|^n)*(a|^k) by A17,BINOM:10;
then a|^l in I by A5,Def3;
then ((n+m) choose r)*a|^l in I by Th17;
hence thesis by A16,Def3;
end;
suppose
l < n;
then ((n+m) + -r) + r < n + r by XREAL_1:6;
then -n + (n+m) < -n + (n + r) by XREAL_1:6;
then consider k being Nat such that
A18: r = m + k by NAT_1:10;
reconsider k as Element of NAT by ORDINAL1:def 12;
b|^r = (b|^m)*(b|^k) by A18,BINOM:10;
then b|^r in I by A8,Def3;
hence thesis by A16,Def3;
end;
end;
A19: now
let j be Element of NAT;
assume that
0 <= j and
A20: j < len p;
thus P[j] implies P[j+1]
proof
assume
A21: f.j in I;
A22: j + 1 <= len p by A20,NAT_1:13;
1 <= j + 1 by NAT_1:11;
then j + 1 in Seg(len p) by A22,FINSEQ_1:1;
then j + 1 in dom p by FINSEQ_1:def 3;
then
A23: p/.(j+1) = p.(j+1) by PARTFUN1:def 6;
then
A24: p/.(j+1) in I by A12,A22,NAT_1:11;
f.(j + 1) = f.j + p/.(j+1) by A11,A20,A23;
hence thesis by A21,A24,Def1;
end;
end;
A25: (a+b)|^(n+m) = Sum((a,b) In_Power (n+m)) by BINOM:25;
A26: P[0] by A10,Th2;
for i being Element of NAT st 0 <= i & i <= len p holds P[i] from
INT_1:sch 7(A26,A19);
then (a+b)|^(n+m) in I by A25,A9;
hence thesis by A3,A6;
end;
hence thesis;
end;
end;
registration
let R be well-unital commutative associative non empty doubleLoopStr, I be
left-ideal non empty Subset of R;
cluster sqrt I -> left-ideal;
coherence
proof
set M ={a where a is Element of R: ex n being Element of NAT st a|^n in I};
M = sqrt I;
then reconsider M as non empty Subset of R;
for y,x being Element of R st x in M holds y*x in M
proof
let y9,x9 be Element of R;
reconsider x = x9,y = y9 as Element of R;
assume x9 in M;
then consider a being Element of R such that
A1: x = a and
A2: ex n being Element of NAT st a|^n in I;
consider n being Element of NAT such that
A3: a|^n in I by A2;
A4: (y*a)|^n = (y|^n)*(a|^n) by BINOM:9;
(y|^n)*(a|^n) in I by A3,Def2;
hence thesis by A1,A4;
end;
hence thesis;
end;
cluster sqrt I -> right-ideal;
coherence;
end;
theorem
for R being well-unital associative non empty doubleLoopStr, I being
non empty Subset of R, a being Element of R holds a in sqrt I iff ex n being
Element of NAT st a|^n in sqrt I
proof
let R be well-unital associative non empty doubleLoopStr, I be non empty
Subset of R, a be Element of R;
A1: now
assume ex n being Element of NAT st a|^n in sqrt I;
then consider n being Element of NAT such that
A2: a|^n in sqrt I;
consider d being Element of R such that
A3: a|^n = d and
A4: ex m being Element of NAT st d|^m in I by A2;
consider m being Element of NAT such that
A5: d|^m in I by A4;
a|^(n*m) = d|^m by A3,BINOM:11;
hence a in sqrt I by A5;
end;
now
A6: a|^1 = a by BINOM:8;
assume a in sqrt I;
hence ex n being Element of NAT st a|^n in sqrt I by A6;
end;
hence thesis by A1;
end;
theorem
for R being left_zeroed right_zeroed add-cancelable distributive
well-unital associative non empty doubleLoopStr, I being add-closed
right-ideal non empty Subset of R, J being add-closed left-ideal non empty
Subset of R holds sqrt (I *' J) = sqrt (I /\ J)
proof
let R be left_zeroed right_zeroed associative add-cancelable distributive
well-unital non empty doubleLoopStr, I be add-closed right-ideal non empty
Subset of R, J be add-closed left-idealnon empty Subset of R;
A1: now
let u be object;
assume u in sqrt (I *' J);
then consider d being Element of R such that
A2: u = d and
A3: ex m being Element of NAT st d|^m in I *' J;
consider m being Element of NAT such that
A4: d|^m in I *' J by A3;
consider s being FinSequence of the carrier of R such that
A5: d|^m = Sum s and
A6: for i being Element of NAT st 1 <= i & i <= len s ex a,b being
Element of R st s.i = a*b & a in I & b in J by A4;
consider f being sequence of the carrier of R such that
A7: Sum s = f.(len s) and
A8: f.0 = 0.R and
A9: for j being Nat, v being Element of R st j < len s & v
= s.(j + 1) holds f.(j + 1) = f.j + v by RLVECT_1:def 12;
defpred P[Element of NAT] means f.$1 in I /\ J;
A10: now
let j be Element of NAT;
assume that
0 <= j and
A11: j < len s;
thus P[j] implies P[j+1]
proof
assume f.j in I /\ J;
then
A12: ex g being Element of R st g = f.j & g in I & g in J;
A13: j + 1 <= len s & 0 + 1 <= j + 1 by A11,NAT_1:13;
then
A14: ex a,b being Element of R st s.(j+1) = a*b & a in I & b in J by A6;
j + 1 in Seg(len s) by A13,FINSEQ_1:1;
then j + 1 in dom s by FINSEQ_1:def 3;
then
A15: s.(j+1) = s/.(j+1) by PARTFUN1:def 6;
then
A16: f.(j+1) = f.j + s/.(j+1) by A9,A11;
s/.(j+1) in J by A15,A14,Def2;
then
A17: f.(j+1) in J by A12,A16,Def1;
s/.(j+1) in I by A15,A14,Def3;
then f.(j+1) in I by A12,A16,Def1;
hence thesis by A17;
end;
end;
f.0 in I & f.0 in J by A8,Th2,Th3;
then
A18: P[0];
for j being Element of NAT st 0 <= j&j<=len s holds P[j] from
INT_1:sch 7(A18,A10);
then Sum s in I /\ J by A7;
hence u in sqrt (I /\ J) by A2,A5;
end;
now
let u be object;
assume u in sqrt (I /\ J);
then consider d being Element of R such that
A19: u = d and
A20: ex m being Element of NAT st d|^m in I /\ J;
consider m being Element of NAT such that
A21: d|^m in I /\ J by A20;
set q = <*d|^m*d|^m*>;
A22: len q = 1 by FINSEQ_1:40;
A23: ex g being Element of R st d|^m = g & g in I & g in J by A21;
A24: for i being Element of NAT st 1 <= i & i <= len q ex x,y being
Element of R st q.i = x*y & x in I & y in J
proof
let i be Element of NAT;
assume
A25: 1 <= i & i <= len q;
then i in Seg(len q) by FINSEQ_1:1;
then i in dom q by FINSEQ_1:def 3;
then
A26: q.i = q/.i by PARTFUN1:def 6;
then q/.i = q.1 by A22,A25,XXREAL_0:1
.= d|^m*d|^m by FINSEQ_1:40;
hence thesis by A23,A26;
end;
d|^(m+m) = d|^m*d|^m by BINOM:10
.= Sum q by BINOM:3;
then d|^(m+m) in I *' J by A24;
hence u in sqrt (I *' J) by A19;
end;
hence thesis by A1,TARSKI:2;
end;
begin :: Noetherian ideals
definition
let L be non empty doubleLoopStr, I be Ideal of L;
attr I is finitely_generated means
ex F being non empty finite Subset of L st I = F-Ideal;
end;
registration
let L be non empty doubleLoopStr;
cluster finitely_generated for Ideal of L;
existence
proof
consider x being object such that
A1: x in the carrier of L by XBOOLE_0:def 1;
reconsider x as Element of L by A1;
take {x}-Ideal;
thus thesis;
end;
end;
registration
let L be non empty doubleLoopStr, F be non empty finite Subset of L;
cluster F-Ideal -> finitely_generated;
coherence;
end;
definition
let L be non empty doubleLoopStr;
attr L is Noetherian means
:Def26:
for I being Ideal of L holds I is finitely_generated;
end;
registration
cluster Euclidian Abelian add-associative right_zeroed right_complementable
well-unital distributive commutative associative non degenerated for
non empty
doubleLoopStr;
existence
proof
take INT.Ring;
thus thesis;
end;
end;
definition
let L be non empty doubleLoopStr;
let I be Ideal of L;
attr I is principal means
ex e being Element of L st I = {e}-Ideal;
end;
definition
let L be non empty doubleLoopStr;
attr L is PID means
for I being Ideal of L holds I is principal;
end;
theorem Th93:
for L being non empty doubleLoopStr, F being non empty Subset of
L st F <> {0.L} ex x being Element of L st x <> 0.L & x in F
proof
let L be non empty doubleLoopStr, F be non empty Subset of L;
assume
A1: F <> {0.L};
now
assume
A2: for x being set st x in F holds x = 0.L;
for x being object holds x in F iff x = 0.L
proof
let e be object;
A3: ex a being object st a in F by XBOOLE_0:def 1;
thus e in F implies e = 0.L by A2;
assume e = 0.L;
hence thesis by A2,A3;
end;
hence contradiction by A1,TARSKI:def 1;
end;
hence thesis;
end;
theorem Th94:
for R being add-associative left_zeroed right_zeroed
right_complementable distributive well-unital Euclidian non empty
doubleLoopStr holds R is PID
proof
let R be add-associative left_zeroed right_zeroed right_complementable
distributive well-unital Euclidian non empty doubleLoopStr;
let I be Ideal of R;
per cases;
suppose
A1: I = {0.R};
set e = 0.R;
take e;
thus thesis by A1,Th44;
end;
suppose
I <> {0.R};
then consider x being Element of R such that
A2: x <> 0.R & x in I by Th93;
set I9 = { y where y is Element of I : y <> 0.R };
A3: I9 c= the carrier of R
proof
let x be object;
assume x in I9;
then ex y being Element of I st x=y & y <> 0.R;
hence thesis;
end;
x in I9 by A2;
then reconsider I9 as non empty Subset of R by A3;
consider f being Function of the carrier of R, NAT such that
A4: for a,b being Element of R st b <> 0.R holds ex q,r being Element
of R st a = q*b + r & (r = 0.R or (f.r qua Element of NAT) < (f.b qua Element
of NAT)) by INT_3:def 8;
set K = the set of all f.i where i is Element of I9 ;
set i = the Element of I9;
A5: f.i in K;
K c= NAT
proof
let x be object;
assume x in K;
then ex e being Element of I9 st f.e = x;
hence thesis;
end;
then reconsider K as non empty Subset of NAT by A5;
set k=min K;
k in K by XXREAL_2:def 7;
then consider e being Element of I9 such that
A6: f.e = k;
e in I9;
then
A7: ex e9 being Element of I st e9=e & e9 <> 0.R;
reconsider e as Element of R;
take e;
now
let x be object;
{e} c= I by A7,ZFMISC_1:31;
then
A8: {e}-Ideal c= I by Def14;
hereby
assume
A9: x in I;
then reconsider x9=x as Element of R;
consider q,r being Element of R such that
A10: x9 = q*e + r and
A11: r = 0.R or (f.r qua Element of NAT) < k by A4,A6,A7;
now
q*e in I by A7,Def2;
then
A12: -(q*e) in I by Th13;
assume
A13: r <> 0.R;
-(q*e) + x9 = -(q*e) + q*e + r by A10,RLVECT_1:def 3
.= 0.R + r by RLVECT_1:5
.= r by ALGSTR_1:def 2;
then r in I by A9,A12,Def1;
then r in I9 by A13;
then f.r in K;
hence contradiction by A11,A13,XXREAL_2:def 7;
end;
then
A14: x9 = q*e by A10,RLVECT_1:def 4;
e in {e} & {e} c= {e}-Ideal by Def14,TARSKI:def 1;
hence x in {e}-Ideal by A14,Def2;
end;
assume x in {e}-Ideal;
hence x in I by A8;
end;
hence thesis by TARSKI:2;
end;
end;
theorem Th95:
for L being non empty doubleLoopStr st L is PID holds L is Noetherian
proof
let L be non empty doubleLoopStr;
assume
A1: L is PID;
let I being Ideal of L;
I is principal by A1;
then consider e being Element of L such that
A2: I = {e}-Ideal;
take {e};
thus thesis by A2;
end;
registration
cluster INT.Ring -> Noetherian;
coherence
proof
INT.Ring is PID by Th94;
hence thesis by Th95;
end;
end;
registration
cluster Noetherian Abelian add-associative right_zeroed right_complementable
well-unital distributive commutative associative non degenerated for
non empty
doubleLoopStr;
existence
proof
take INT.Ring;
thus thesis;
end;
end;
theorem :: Lemma_4_5_i_ii:
for R being Noetherian add-associative left_zeroed right_zeroed
add-cancelable associative distributive well-unital non empty doubleLoopStr
for B being non empty Subset of R ex C being non empty finite Subset of R st C
c= B & C-Ideal = B-Ideal
proof
let R be Noetherian add-associative left_zeroed right_zeroed add-cancelable
associative distributive well-unital non empty doubleLoopStr;
let B be non empty Subset of R;
defpred P[object,object] means
ex cL being non empty LinearCombination of B st $1
= Sum cL & ex fsB being FinSequence of B st dom fsB = dom cL & $2 = rng fsB &
for i being Nat st i in dom cL ex u, v being Element of R st cL/.i =
u*((fsB/.i qua Element of B) qua Element of R)*v;
B-Ideal is finitely_generated by Def26;
then consider D being non empty finite Subset of R such that
A1: D-Ideal = B-Ideal;
A2: D c= B-Ideal by A1,Def14;
A3: for e being object st e in D ex u being object st u in bool B & P[e,u]
proof
let e be object;
assume e in D;
then consider cL being LinearCombination of B such that
A4: e = Sum cL by A2,Th60;
per cases;
suppose
A5: cL is non empty;
defpred P1[set,Element of B] means ex u, v being Element of R st cL/.$1
= u*($2 qua Element of R)*v;
A6: now
let k be Nat;
assume k in Seg len cL;
then k in dom cL by FINSEQ_1:def 3;
then consider u, v being Element of R, a being Element of B such that
A7: cL/.k = u*a*v by Def8;
take d = a;
thus P1[k,d] by A7;
end;
consider fsB being FinSequence of B such that
A8: dom fsB = Seg len cL & for k being Nat st k in Seg
len cL holds P1[k,(fsB/.k) qua Element of B] from RECDEF_1:sch 17(A6);
take u = rng fsB;
thus u in bool B;
dom cL = Seg len cL by FINSEQ_1:def 3;
hence thesis by A4,A5,A8;
end;
suppose
A9: cL is empty;
set b = the Element of B;
set kL = <*0.R*b*0.R*>;
now
let i be set;
assume
A10: i in dom kL;
take u = 0.R, v = 0.R, b;
i in Seg len kL by A10,FINSEQ_1:def 3;
then i in {1} by FINSEQ_1:2,40;
then i = 1 by TARSKI:def 1;
hence kL/.i = u*b*v by FINSEQ_4:16;
end;
then reconsider kL as non empty LinearCombination of B by Def8;
cL = <*>the carrier of R by A9;
then
A11: e = 0.R by A4,RLVECT_1:43
.= 0.R*b*0.R by BINOM:2
.= Sum kL by BINOM:3;
defpred P2[Nat,Element of B] means ex u, v being Element of R
st kL/.$1 = u*(($2 qua Element of B) qua Element of R)*v;
A12: now
let k be Nat;
assume
A13: k in Seg len kL;
take b;
k in {1} by A13,FINSEQ_1:2,40;
then k = 1 by TARSKI:def 1;
hence P2[k,b] by FINSEQ_4:16;
end;
consider fsB being FinSequence of B such that
A14: dom fsB = Seg len kL & for k being Nat st k in Seg
len kL holds P2[k,(fsB/.k) qua Element of B] from RECDEF_1:sch 17(A12);
take u = rng fsB;
thus u in bool B;
dom kL = Seg len kL by FINSEQ_1:def 3;
hence thesis by A11,A14;
end;
end;
consider f being Function of D, bool B such that
A15: for e being object st e in D holds P[e,f.e] from FUNCT_2:sch 1(A3);
A16: now
let r be set;
assume r in rng f;
then consider x being object such that
A17: x in dom f & r = f.x by FUNCT_1:def 3;
ex cL being non empty LinearCombination of B st x = Sum cL & ex fsB
being FinSequence of B st dom fsB = dom cL & r = rng fsB &
for i being Nat st i in dom cL ex u, v being Element of R
st cL/.i = u*(( fsB/.i qua
Element of B) qua Element of R)*v by A15,A17;
hence r is finite;
end;
reconsider rf = rng f as Subset-Family of B;
reconsider C = union rf as Subset of B;
consider r being object such that
A18: r in rng f by XBOOLE_0:def 1;
consider x being object such that
A19: x in dom f & r = f.x by A18,FUNCT_1:def 3;
reconsider r as set by TARSKI:1;
ex cL being non empty LinearCombination of B st x = Sum cL & ex fsB
being FinSequence of B st dom fsB = dom cL & r = rng fsB &
for i being Nat st i in dom cL ex u, v being Element of R
st cL/.i = u*(( fsB/.i qua
Element of B) qua Element of R)*v by A15,A19;
then r is non empty by RELAT_1:42;
then ex x being object st x in r;
then reconsider C as non empty finite Subset of R by A18,A16,FINSET_1:7
,TARSKI:def 4,XBOOLE_1:1;
now
let d be object;
assume
A20: d in D;
then consider cL being non empty LinearCombination of B such that
A21: d = Sum cL and
A22: ex fsB being FinSequence of B st dom fsB = dom cL & f.d = rng fsB
& for i being Nat st i in dom cL ex u, v being Element of R st cL/.i
= u*((fsB/.i qua Element of B) qua Element of R)*v by A15;
d in dom f by A20,FUNCT_2:def 1;
then f.d in rng f by FUNCT_1:def 3;
then
A23: f.d c= C by ZFMISC_1:74;
now
let i be set;
consider fsB being FinSequence of B such that
A24: dom fsB = dom cL and
A25: f.d = rng fsB and
A26: for i being Nat st i in dom cL ex u, v being Element
of R st cL/.i = u*((fsB/.i qua Element of B) qua Element of R)*v by A22;
assume
A27: i in dom cL;
then fsB/.i = fsB.i by A24,PARTFUN1:def 6;
then
A28: fsB/.i in f.d by A27,A24,A25,FUNCT_1:def 3;
ex u, v being Element of R st cL/.i = u*((fsB/.i qua Element of B)
qua Element of R)*v by A27,A26;
hence
ex u,v being Element of R, a being Element of C st cL/.i = u*a*v by A23
,A28;
end;
then reconsider cL9= cL as LinearCombination of C by Def8;
d = Sum cL9 by A21;
hence d in C-Ideal by Th60;
end;
then D c= C-Ideal;
then D-Ideal c= (C-Ideal)-Ideal by Th57;
then
A29: B-Ideal c= C-Ideal by A1,Th44;
take C;
C-Ideal c= B-Ideal by Th57;
hence thesis by A29;
end;
theorem :: Lemma_4_5_ii_iii:
for R being non empty doubleLoopStr st for B being non empty Subset
of R ex C being non empty finite Subset of R st C c= B & C-Ideal = B-Ideal for
a being sequence of R ex m being Element of NAT st a.(m+1) in (rng (a|(m+1)))
-Ideal
proof
let R be non empty doubleLoopStr;
assume
A1: for B being non empty Subset of R ex C being non empty finite Subset
of R st C c= B & C-Ideal = B-Ideal;
let a being sequence of R;
reconsider B = rng a as non empty Subset of R;
consider C being non empty finite Subset of R such that
A2: C c= B and
A3: C-Ideal = B-Ideal by A1;
defpred P[object,object] means $1 = a.$2;
A4: dom a = NAT by FUNCT_2:def 1;
A5: for e being object st e in C ex u being object st u in NAT & P[e,u]
proof
let e be object;
assume e in C;
then consider u being object such that
A6: u in dom a and
A7: e = a.u by A2,FUNCT_1:def 3;
take u;
thus u in NAT by A6;
thus thesis by A7;
end;
consider f being Function of C, NAT such that
A8: for e being object st e in C holds P[e,f.e] from FUNCT_2:sch 1(A5);
set Rf = rng f;
reconsider Rf as non empty finite Subset of NAT;
reconsider m = max Rf as Element of NAT by ORDINAL1:def 12;
set D = rng (a | Segm(m+1));
A9: dom f = C by FUNCT_2:def 1;
A10: C c= D
proof
let X be object;
set fx = f.X;
assume
A11: X in C;
then f.X in Rf by A9,FUNCT_1:def 3;
then fx <= m by XXREAL_2:def 8;
then fx < m+1 by NAT_1:13;
then fx in Segm(m+1) by NAT_1:44;
then a.fx in rng (a | Segm(m+1)) by A4,FUNCT_1:50;
hence thesis by A8,A11;
end;
then reconsider D as non empty Subset of R;
A12: D-Ideal c= B-Ideal by Th57,RELAT_1:70;
B-Ideal c= D-Ideal by A3,A10,Th57;
then
A13: D-Ideal = B-Ideal by A12;
take m;
B c= B-Ideal & a.(m+1) in B by Def14,FUNCT_2:4;
hence thesis by A13;
end;
registration
let X, Y be non empty set, f be Function of X, Y, A be non empty Subset of X;
cluster f|A -> non empty;
coherence
proof
dom f = X by FUNCT_2:def 1;
then (dom f) /\ A is non empty by XBOOLE_1:28;
then (dom f) meets A;
hence thesis by RELAT_1:66;
end;
end;
theorem :: Lemma_4_5_iii_iv:
for R being non empty doubleLoopStr st for a being sequence of R ex
m being Element of NAT st a.(m+1) in (rng (a|(m+1)))-Ideal holds not ex F being
sequence of bool (the carrier of R) st (for i being Nat holds F
.i is Ideal of R) & (for j,k being Nat st j < k holds F.j c< F.k)
proof
let R being non empty doubleLoopStr;
assume
A1: for a being sequence of R ex m being Element of NAT st a.(m+1) in (
rng (a|(m+1)))-Ideal;
given F being sequence of bool (the carrier of R) such that
A2: for i being Nat holds F.i is Ideal of R and
A3: for j,k being Nat st j < k holds F.j c< F.k;
defpred P[object,object] means
ex n being Element of NAT st n = $1 & (n = 0
implies $2 in F.0) & (n > 0 implies ex k being Element of NAT st n = k+1 & $2
in F.n \ F.k);
A4: for e being object st e in NAT
ex u being object st u in the carrier of R & P[e,u]
proof
let e be object;
assume e in NAT;
then reconsider n=e as Element of NAT;
per cases;
suppose
A5: n = 0;
F.0 is Ideal of R by A2;
then consider u being object such that
A6: u in F.0 by XBOOLE_0:def 1;
take u;
thus u in the carrier of R by A6;
take n;
thus n=e;
thus thesis by A5,A6;
end;
suppose
n > 0;
then consider k being Nat such that
A7: n=k+1 by NAT_1:6;
reconsider k as Element of NAT by ORDINAL1:def 12;
n > k by A7,NAT_1:13;
then not F.n c= F.k by A3,XBOOLE_1:60;
then F.n \ F.k is non empty by XBOOLE_1:37;
then consider u being object such that
A8: u in F.n \ F.k;
take u;
thus u in the carrier of R by A8;
take n;
thus n=e;
thus thesis by A7,A8;
end;
end;
consider f being sequence of the carrier of R such that
A9: for e being object st e in NAT holds P[e, f.e] from FUNCT_2:sch 1(A4);
consider m being Element of NAT such that
A10: f.(m+1) in (rng (f|(m+1)))-Ideal by A1;
reconsider m1 = m+1 as non zero Nat;
A11: ex n being Element of NAT st n = m+1 &( n = 0 implies f.( m+1) in F.0)&(
n > 0 implies ex k being Element of NAT st n = k+1 & f.(m+1) in F.n \ F.k)
by A9;
defpred P[Nat] means rng (f|Segm($1+1)) c= F.$1;
A12: for k being Nat st P[k] holds P[k+1]
proof
let k be Nat such that
A13: rng (f|Segm(k+1)) c= F.k;
let y be object;
assume y in rng (f|Segm((k+1)+1));
then consider x being object such that
A14: x in dom (f|Segm((k+1)+1)) and
A15: y = (f|Segm((k+1)+1)).x by FUNCT_1:def 3;
A16: x in dom f by A14,RELAT_1:57;
reconsider nx = x as Element of NAT by A14;
x in Segm((k+1)+1) by A14,RELAT_1:57;
then nx < (k+1)+1 by NAT_1:44;
then
A17: nx <= k+1 by NAT_1:13;
per cases by A17,XXREAL_0:1;
suppose
nx < k+1;
then
A18: nx in Segm(k+1) by NAT_1:44;
k < k+1 by NAT_1:13;
then F.k c< F.(k+1) by A3;
then
A19: F.k c= F.(k+1);
y = f.nx by A14,A15,FUNCT_1:47;
then y in rng(f|Segm(k+1)) by A16,A18,FUNCT_1:50;
then y in F.k by A13;
hence thesis by A19;
end;
suppose
A20: nx = k+1;
y = f.nx & ex n being Element of NAT st n = nx &( n = 0 implies f.
nx in F.0 )&( n > 0 implies ex k being Element of NAT st n = k+1 & f.nx in F.n
\ F. k) by A9,A14,A15,FUNCT_1:47;
hence thesis by A20,XBOOLE_0:def 5;
end;
end;
F.m is Ideal of R by A2;
then
A21: F.m = (F.m)-Ideal by Th44;
A22: P[0]
proof
let y be object;
assume y in rng (f|Segm(0+1));
then consider x being object such that
A23: x in dom (f|Segm(1)) and
A24: y = (f|Segm(1)).x by FUNCT_1:def 3;
x in Segm(1) & ex n being Element of NAT st n = x & (n = 0 implies f.
x in F. 0) & (n > 0 implies ex k being Element of NAT st n = k+1 & f.x in F.n \
F.k) by A9,A23,RELAT_1:57;
hence thesis by A24,CARD_1:49,FUNCT_1:49,TARSKI:def 1;
end;
for m being Nat holds P[m] from NAT_1:sch 2(A22,A12);
then (rng (f|Segm m1))-Ideal c= F.m by A21,Th57;
hence contradiction by A10,A11,XBOOLE_0:def 5;
end;
theorem :: Lemma_4_5_iv_i:
for R being non empty doubleLoopStr st not ex F being sequence of
bool (the carrier of R) st (for i being Element of NAT holds F.i is Ideal
of R) & (for j,k being Element of NAT st j < k holds F.j c< F.k) holds R is
Noetherian
proof
let R being non empty doubleLoopStr such that
A1: not ex F being sequence of bool (the carrier of R) st (for i
being Element of NAT holds F.i is Ideal of R) & (for j,k being Element of NAT
st j < k holds F.j c< F.k) and
A2: not R is Noetherian;
consider I being Ideal of R such that
A3: I is not finitely_generated by A2;
set D = { S where S is Subset of R : S is non empty finite Subset of I };
consider e being object such that
A4: e in I by XBOOLE_0:def 1;
reconsider e as Element of R by A4;
{e} c= I by A4,ZFMISC_1:31;
then
A5: {e} in D;
D c= bool the carrier of R
proof
let x be object;
assume x in D;
then
ex s being Subset of R st x = s & s is non empty finite Subset of I;
hence thesis;
end;
then reconsider D as non empty Subset-Family of R by A5;
reconsider e9={e} as Element of D by A5;
defpred P[set,Element of D,set] means ex r being Element of R st r in I \ $2
-Ideal & $3 = $2 \/ {r};
A6: for n be Nat for x be Element of D ex y be Element of D st P
[n,x,y]
proof
let n be Nat, x be Element of D;
x in D;
then consider x9 being Subset of R such that
A7: x9 = x and
A8: x9 is non empty finite Subset of I;
reconsider x19=x9 as non empty finite Subset of I by A8;
x9-Ideal c= I-Ideal by A8,Th57;
then x9-Ideal c= I by Th44;
then not I c= x9-Ideal by A3,A8,XBOOLE_0:def 10;
then consider r being object such that
A9: r in I and
A10: not r in x9-Ideal;
set y=x19 \/ {r};
A11: y c= I
proof
let x be object;
assume x in y;
then x in x19 or x in {r} by XBOOLE_0:def 3;
hence thesis by A9,TARSKI:def 1;
end;
then y is Subset of R by XBOOLE_1:1;
then
A12: y in D by A11;
reconsider r as Element of R by A9;
reconsider y as Element of D by A12;
take y;
take r;
thus thesis by A7,A9,A10,XBOOLE_0:def 5;
end;
consider f be sequence of D such that
A13: f.0 = e9 & for n be Nat holds P[n,(f.n) qua Element of D
,f.(n+1)] from RECDEF_1:sch 2(A6);
defpred Q[Nat,Subset of R] means ex c being Subset of R st c = f.
$1 & $2 = c-Ideal;
A14: for x being Element of NAT ex y being Subset of R st Q[x,y]
proof
let x be Element of NAT;
f.x in D;
then consider c being Subset of R such that
A15: c = f.x and
c is non empty finite Subset of I;
reconsider y = c-Ideal as Subset of R;
take y;
take c;
thus thesis by A15;
end;
consider F being sequence of bool the carrier of R such that
A16: for x being Element of NAT holds Q[x,F.x] from FUNCT_2:sch 3(A14);
A17: for x being Nat holds Q[x,F.x]
proof let x be Nat;
x in NAT by ORDINAL1:def 12;
hence thesis by A16;
end;
A18: for j,k being Element of NAT st j < k holds F.j c< F.k
proof
let j,k be Element of NAT;
defpred P[Nat] means F.j c< F.(j+1+$1);
assume j < k;
then j+1 <= k by NAT_1:13;
then consider i being Nat such that
A19: k = j + 1 + i by NAT_1:10;
A20: for i being Nat holds F.i c< F.(i+1)
proof
let i be Nat;
consider c being Subset of R such that
A21: c = f.i and
A22: F.i = c-Ideal by A17;
consider c1 being Subset of R such that
A23: c1 = f.(i+1) and
A24: F.(i+1) = c1-Ideal by A17;
A25: c1 c= c1-Ideal by Def14;
consider r being Element of R such that
A26: r in I \ c-Ideal and
A27: c1 = c \/ {r} by A13,A21,A23;
c in D by A21;
then
ex c9 being Subset of R st c9 = c & c9 is non empty finite Subset of I;
hence F.i c= F.(i+1) by A22,A24,A27,Th57,XBOOLE_1:7;
r in {r} by TARSKI:def 1;
then r in c1 by A27,XBOOLE_0:def 3;
hence thesis by A22,A24,A26,A25,XBOOLE_0:def 5;
end;
A28: for i being Nat st P[i] holds P[i+1]
proof
let i be Nat such that
A29: F.j c= F.(j+1+i) and
F.j <> F.(j+1+i);
A30: F.(j+1+i) c< F.((j+1+i)+1) by A20;
then F.(j+1+i) c= F.((j+1+i)+1);
hence F.j c= F.(j+1+(i+1)) by A29;
thus thesis by A29,A30,XBOOLE_0:def 8;
end;
A31: P[0] by A20;
A32: for i being Nat holds P[i] from NAT_1:sch 2(A31, A28);
thus thesis by A32,A19;
end;
for i being Element of NAT holds F.i is Ideal of R
proof
let i be Element of NAT;
ex c being Subset of R st c = f.i & F.i = c-Ideal by A17;
hence thesis;
end;
hence contradiction by A1,A18;
end;