let f be PartFunc of (REAL 3),REAL; for u0 being Element of REAL 3
for N being Neighbourhood of (proj (1,3)) . u0 st f is_partial_differentiable_in u0,1 & N c= dom (SVF1 (1,f,u0)) holds
for h being non-zero 0 -convergent Real_Sequence
for c being constant Real_Sequence st rng c = {((proj (1,3)) . u0)} & rng (h + c) c= N holds
( (h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c)) is convergent & partdiff (f,u0,1) = lim ((h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c))) )
let u0 be Element of REAL 3; for N being Neighbourhood of (proj (1,3)) . u0 st f is_partial_differentiable_in u0,1 & N c= dom (SVF1 (1,f,u0)) holds
for h being non-zero 0 -convergent Real_Sequence
for c being constant Real_Sequence st rng c = {((proj (1,3)) . u0)} & rng (h + c) c= N holds
( (h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c)) is convergent & partdiff (f,u0,1) = lim ((h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c))) )
let N be Neighbourhood of (proj (1,3)) . u0; ( f is_partial_differentiable_in u0,1 & N c= dom (SVF1 (1,f,u0)) implies for h being non-zero 0 -convergent Real_Sequence
for c being constant Real_Sequence st rng c = {((proj (1,3)) . u0)} & rng (h + c) c= N holds
( (h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c)) is convergent & partdiff (f,u0,1) = lim ((h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c))) ) )
assume A1:
( f is_partial_differentiable_in u0,1 & N c= dom (SVF1 (1,f,u0)) )
; for h being non-zero 0 -convergent Real_Sequence
for c being constant Real_Sequence st rng c = {((proj (1,3)) . u0)} & rng (h + c) c= N holds
( (h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c)) is convergent & partdiff (f,u0,1) = lim ((h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c))) )
let h be non-zero 0 -convergent Real_Sequence; for c being constant Real_Sequence st rng c = {((proj (1,3)) . u0)} & rng (h + c) c= N holds
( (h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c)) is convergent & partdiff (f,u0,1) = lim ((h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c))) )
let c be constant Real_Sequence; ( rng c = {((proj (1,3)) . u0)} & rng (h + c) c= N implies ( (h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c)) is convergent & partdiff (f,u0,1) = lim ((h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c))) ) )
assume A2:
( rng c = {((proj (1,3)) . u0)} & rng (h + c) c= N )
; ( (h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c)) is convergent & partdiff (f,u0,1) = lim ((h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c))) )
consider x0, y0, z0 being Real such that
A3:
( u0 = <*x0,y0,z0*> & ex N1 being Neighbourhood of x0 st
( N1 c= dom (SVF1 (1,f,u0)) & ex L being LinearFunc ex R being RestFunc st
for x being Real st x in N1 holds
((SVF1 (1,f,u0)) . x) - ((SVF1 (1,f,u0)) . x0) = (L . (x - x0)) + (R . (x - x0)) ) )
by A1, Th13;
consider N1 being Neighbourhood of x0 such that
A4:
( N1 c= dom (SVF1 (1,f,u0)) & ex L being LinearFunc ex R being RestFunc st
for x being Real st x in N1 holds
((SVF1 (1,f,u0)) . x) - ((SVF1 (1,f,u0)) . x0) = (L . (x - x0)) + (R . (x - x0)) )
by A3;
consider L being LinearFunc, R being RestFunc such that
A5:
for x being Real st x in N1 holds
((SVF1 (1,f,u0)) . x) - ((SVF1 (1,f,u0)) . x0) = (L . (x - x0)) + (R . (x - x0))
by A4;
A6:
(proj (1,3)) . u0 = x0
by A3, Th1;
then consider N2 being Neighbourhood of x0 such that
A7:
( N2 c= N & N2 c= N1 )
by RCOMP_1:17;
consider g being Real such that
A8:
( 0 < g & N2 = ].(x0 - g),(x0 + g).[ )
by RCOMP_1:def 6;
A9:
x0 in N2
ex n being Element of NAT st
( rng (c ^\ n) c= N2 & rng ((h + c) ^\ n) c= N2 )
proof
x0 in rng c
by A2, A6, TARSKI:def 1;
then A11:
lim c = x0
by SEQ_4:25;
(
h is
convergent &
lim h = 0 )
;
then A12:
lim (h + c) =
0 + x0
by A11, SEQ_2:6
.=
x0
;
consider n being
Nat such that A13:
for
m being
Nat st
n <= m holds
|.(((h + c) . m) - x0).| < g
by A8, A12, SEQ_2:def 7;
A14:
rng (c ^\ n) = {x0}
by A2, A6, VALUED_0:26;
reconsider n =
n as
Element of
NAT by ORDINAL1:def 12;
take
n
;
( rng (c ^\ n) c= N2 & rng ((h + c) ^\ n) c= N2 )
thus
rng (c ^\ n) c= N2
by A9, A14, TARSKI:def 1;
rng ((h + c) ^\ n) c= N2
let y be
object ;
TARSKI:def 3 ( not y in rng ((h + c) ^\ n) or y in N2 )
assume
y in rng ((h + c) ^\ n)
;
y in N2
then consider m being
Element of
NAT such that A15:
y = ((h + c) ^\ n) . m
by FUNCT_2:113;
n + 0 <= n + m
by XREAL_1:7;
then
|.(((h + c) . (n + m)) - x0).| < g
by A13;
then
(
- g < ((h + c) . (m + n)) - x0 &
((h + c) . (m + n)) - x0 < g )
by SEQ_2:1;
then
(
- g < (((h + c) ^\ n) . m) - x0 &
(((h + c) ^\ n) . m) - x0 < g )
by NAT_1:def 3;
then
(
x0 + (- g) < ((h + c) ^\ n) . m &
((h + c) ^\ n) . m < x0 + g )
by XREAL_1:19, XREAL_1:20;
hence
y in N2
by A8, A15;
verum
end;
then consider n being Element of NAT such that
A16:
( rng (c ^\ n) c= N2 & rng ((h + c) ^\ n) c= N2 )
;
A17:
rng (c ^\ n) c= dom (SVF1 (1,f,u0))
proof
let y be
object ;
TARSKI:def 3 ( not y in rng (c ^\ n) or y in dom (SVF1 (1,f,u0)) )
A18:
rng (c ^\ n) = rng c
by VALUED_0:26;
assume
y in rng (c ^\ n)
;
y in dom (SVF1 (1,f,u0))
then
y = x0
by A2, A6, A18, TARSKI:def 1;
then
y in N
by A7, A9;
hence
y in dom (SVF1 (1,f,u0))
by A1;
verum
end;
A19:
rng ((h + c) ^\ n) c= dom (SVF1 (1,f,u0))
by A16, A7, A1;
A20:
rng c c= dom (SVF1 (1,f,u0))
A21:
rng (h + c) c= dom (SVF1 (1,f,u0))
by A2, A1;
A22:
for x being Real st x in N2 holds
((SVF1 (1,f,u0)) . x) - ((SVF1 (1,f,u0)) . x0) = (L . (x - x0)) + (R . (x - x0))
by A5, A7;
A23:
for k being Element of NAT holds ((SVF1 (1,f,u0)) . (((h + c) ^\ n) . k)) - ((SVF1 (1,f,u0)) . ((c ^\ n) . k)) = (L . ((h ^\ n) . k)) + (R . ((h ^\ n) . k))
proof
let k be
Element of
NAT ;
((SVF1 (1,f,u0)) . (((h + c) ^\ n) . k)) - ((SVF1 (1,f,u0)) . ((c ^\ n) . k)) = (L . ((h ^\ n) . k)) + (R . ((h ^\ n) . k))
((h + c) ^\ n) . k in rng ((h + c) ^\ n)
by VALUED_0:28;
then A24:
((h + c) ^\ n) . k in N2
by A16;
A25:
(((h + c) ^\ n) . k) - ((c ^\ n) . k) =
(((h ^\ n) + (c ^\ n)) . k) - ((c ^\ n) . k)
by SEQM_3:15
.=
(((h ^\ n) . k) + ((c ^\ n) . k)) - ((c ^\ n) . k)
by SEQ_1:7
.=
(h ^\ n) . k
;
A26:
(c ^\ n) . k in rng (c ^\ n)
by VALUED_0:28;
rng (c ^\ n) = rng c
by VALUED_0:26;
then
(c ^\ n) . k = x0
by A2, A6, A26, TARSKI:def 1;
hence
((SVF1 (1,f,u0)) . (((h + c) ^\ n) . k)) - ((SVF1 (1,f,u0)) . ((c ^\ n) . k)) = (L . ((h ^\ n) . k)) + (R . ((h ^\ n) . k))
by A5, A7, A24, A25;
verum
end;
A27:
L is total
by FDIFF_1:def 3;
A28:
R is total
by FDIFF_1:def 2;
A29:
((SVF1 (1,f,u0)) /* ((h + c) ^\ n)) - ((SVF1 (1,f,u0)) /* (c ^\ n)) = (L /* (h ^\ n)) + (R /* (h ^\ n))
proof
now for k being Element of NAT holds (((SVF1 (1,f,u0)) /* ((h + c) ^\ n)) - ((SVF1 (1,f,u0)) /* (c ^\ n))) . k = ((L /* (h ^\ n)) + (R /* (h ^\ n))) . klet k be
Element of
NAT ;
(((SVF1 (1,f,u0)) /* ((h + c) ^\ n)) - ((SVF1 (1,f,u0)) /* (c ^\ n))) . k = ((L /* (h ^\ n)) + (R /* (h ^\ n))) . kthus (((SVF1 (1,f,u0)) /* ((h + c) ^\ n)) - ((SVF1 (1,f,u0)) /* (c ^\ n))) . k =
(((SVF1 (1,f,u0)) /* ((h + c) ^\ n)) . k) - (((SVF1 (1,f,u0)) /* (c ^\ n)) . k)
by RFUNCT_2:1
.=
((SVF1 (1,f,u0)) . (((h + c) ^\ n) . k)) - (((SVF1 (1,f,u0)) /* (c ^\ n)) . k)
by A19, FUNCT_2:108
.=
((SVF1 (1,f,u0)) . (((h + c) ^\ n) . k)) - ((SVF1 (1,f,u0)) . ((c ^\ n) . k))
by A17, FUNCT_2:108
.=
(L . ((h ^\ n) . k)) + (R . ((h ^\ n) . k))
by A23
.=
((L /* (h ^\ n)) . k) + (R . ((h ^\ n) . k))
by A27, FUNCT_2:115
.=
((L /* (h ^\ n)) . k) + ((R /* (h ^\ n)) . k)
by A28, FUNCT_2:115
.=
((L /* (h ^\ n)) + (R /* (h ^\ n))) . k
by SEQ_1:7
;
verum end;
hence
((SVF1 (1,f,u0)) /* ((h + c) ^\ n)) - ((SVF1 (1,f,u0)) /* (c ^\ n)) = (L /* (h ^\ n)) + (R /* (h ^\ n))
by FUNCT_2:63;
verum
end;
A30:
( ((L /* (h ^\ n)) + (R /* (h ^\ n))) (#) ((h ^\ n) ") is convergent & lim (((L /* (h ^\ n)) + (R /* (h ^\ n))) (#) ((h ^\ n) ")) = L . 1 )
A40:
N2 c= dom (SVF1 (1,f,u0))
by A1, A7;
A41: ((L /* (h ^\ n)) + (R /* (h ^\ n))) (#) ((h ^\ n) ") =
((((SVF1 (1,f,u0)) /* (h + c)) ^\ n) - ((SVF1 (1,f,u0)) /* (c ^\ n))) (#) ((h ^\ n) ")
by A21, A29, VALUED_0:27
.=
((((SVF1 (1,f,u0)) /* (h + c)) ^\ n) - (((SVF1 (1,f,u0)) /* c) ^\ n)) (#) ((h ^\ n) ")
by A20, VALUED_0:27
.=
((((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c)) ^\ n) (#) ((h ^\ n) ")
by SEQM_3:17
.=
((((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c)) ^\ n) (#) ((h ") ^\ n)
by SEQM_3:18
.=
((((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c)) (#) (h ")) ^\ n
by SEQM_3:19
;
then A42:
L . 1 = lim ((h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c)))
by A30, SEQ_4:22;
thus
(h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c)) is convergent
by A30, A41, SEQ_4:21; partdiff (f,u0,1) = lim ((h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c)))
thus
partdiff (f,u0,1) = lim ((h ") (#) (((SVF1 (1,f,u0)) /* (h + c)) - ((SVF1 (1,f,u0)) /* c)))
by A1, A3, A22, A40, A42, Th16; verum