The Mizar article:

Lim-Inf Convergence

by
Bartlomiej Skorulski

Received January 6, 2000

Copyright (c) 2000 Association of Mizar Users

MML identifier: WAYBEL28
[ MML identifier index ]


environ

 vocabulary BHSP_3, WAYBEL_0, ORDINAL2, YELLOW_0, FUNCT_1, RELAT_1, LATTICES,
      YELLOW_2, LATTICE3, YELLOW_6, ORDERS_1, PRE_TOPC, RELAT_2, QUANTAL1,
      SUBSET_1, SEQM_3, SETFAM_1, WAYBEL11, CLASSES1, CAT_1, PROB_1, YELLOW_9,
      BOOLE, WAYBEL28;
 notation TARSKI, XBOOLE_0, ZFMISC_1, SUBSET_1, STRUCT_0, RELAT_1, RELSET_1,
      FUNCT_1, FUNCT_2, CLASSES1, SEQM_3, GRCAT_1, PRE_TOPC, ORDERS_1,
      LATTICE3, YELLOW_0, WAYBEL_0, YELLOW_2, WAYBEL_3, YELLOW_6, YELLOW_9,
      WAYBEL_9, WAYBEL11, WAYBEL17;
 constructors SEQM_3, GRCAT_1, TOPS_2, CLASSES1, YELLOW_2, WAYBEL_3, YELLOW_9,
      WAYBEL17, MEMBERED, PARTFUN1;
 clusters STRUCT_0, SUBSET_1, RELSET_1, LATTICE3, WAYBEL_0, WAYBEL_3, YELLOW_6,
      YELLOW_9, WAYBEL11, WAYBEL17, CLASSES2, MEMBERED, ZFMISC_1, PARTFUN1;
 requirements SUBSET, BOOLE;
 definitions TARSKI, XBOOLE_0, RELAT_1, YELLOW_6, WAYBEL11;
 theorems ZFMISC_1, RELAT_1, FUNCT_1, FUNCT_2, SEQM_3, GRCAT_1, PRE_TOPC,
      TOPS_1, ORDERS_1, CLASSES1, WAYBEL_0, YELLOW_0, LATTICE3, YELLOW_2,
      YELLOW_4, YELLOW_6, WAYBEL_8, YELLOW_9, WAYBEL_9, WAYBEL11, YELLOW12,
      WAYBEL17, WAYBEL21, XBOOLE_1;
 schemes RELSET_1, COMPTS_1;

begin

theorem Th1:
for L being complete LATTICE, N being net of L holds
  inf N <= lim_inf N
proof
  let L be complete LATTICE, N be net of L;
  consider j being Element of N;
A1:
  ex_inf_of {N.i where i is Element of N:i >= j},L &
    ex_inf_of rng the mapping of N,L by YELLOW_0:17;
    {N.i where i is Element of N:i >= j} c= rng the mapping of N
  proof
    let x be set;
    assume x in {N.i where i is Element of N:i >= j};
    then consider i being Element of N such that
A2:   x = N.i and
        i >= j;
A3: x = (the mapping of N).i by A2,WAYBEL_0:def 8;
      dom(the mapping of N) = the carrier of N by FUNCT_2:def 1;
    hence x in rng(the mapping of N) by A3,FUNCT_1:def 5;
  end;
  then A4: "/\"({N.i where i is Element of N:i >= j},L) >=
      "/\"(rng the mapping of N,L) by A1,YELLOW_0:35;
  set x = "/\"({N.i where i is Element of N:i >= j},L);
    x >= Inf(the mapping of N) by A4,YELLOW_2:def 6;
  then A5: inf N <= x by WAYBEL_9:def 2;
  set X = {"/\"({N.i where i is Element of N :i >= j1},L)
    where j1 is Element of N: not contradiction};
    X c= the carrier of L
  proof
    let x be set;
    assume x in X;
    then consider j1 being Element of N such that
A6:   x = "/\"({N.i where i is Element of N :i >= j1},L);
    thus x in the carrier of L by A6;
  end;
  then reconsider X as Subset of L;
  reconsider X as Subset of L;
A7:
  x in X;
    ex_sup_of X,L by YELLOW_0:17;
  then X is_<=_than "\/"(X,L) by YELLOW_0:def 9;
  then x <= "\/"(X,L) by A7,LATTICE3:def 9;
  then inf N <= "\/"(X,L) by A5,YELLOW_0:def 2;
  hence inf N <= lim_inf N by WAYBEL11:def 6;
end;

theorem ::3.1 Proposition (1)=>(2)
  for L being complete LATTICE, N being net of L, x being Element of L holds
  (for M being subnet of N holds x = lim_inf M) implies
    (x=lim_inf N & for M being subnet of N holds x >= inf M)
proof
  let L be complete LATTICE, N be net of L, x be Element of L;
  assume
A1:  for M being subnet of N holds x = lim_inf M;
    N is subnet of N by YELLOW_6:23;
  hence x=lim_inf N by A1;
  let M be subnet of N;
    x = lim_inf M by A1;
  hence x >= inf M by Th1;
end;

theorem ::3.1 Proposition (1b)=>(2b)
Th3:
for L being complete LATTICE, N being net of L, x being Element of L
    st N in NetUniv L
holds
  (for M being subnet of N st M in NetUniv L holds x = lim_inf M) implies
    (x=lim_inf N & for M being subnet of N st M in NetUniv L holds x >= inf M)
proof
  let L be complete LATTICE, N be net of L, x be Element of L;
  assume
A1:
  N in NetUniv L;
  assume
A2:  for M being subnet of N st M in NetUniv L holds x = lim_inf M;
    N is subnet of N by YELLOW_6:23;
  hence x=lim_inf N by A1,A2;
  let M be subnet of N;
  assume
      M in NetUniv L;
  then x = lim_inf M by A2;
  hence x >= inf M by Th1;
end;

definition
  let N be non empty RelStr;
  let f be map of N,N;
  attr f is greater_or_equal_to_id means :Def1:
    for j being Element of N holds j <= f.j;
end;

theorem Th4:
for N being reflexive non empty RelStr holds id N is greater_or_equal_to_id
proof
  let N be reflexive non empty RelStr;
  let j be Element of N;
  reconsider n=j as Element of N;
    n = (id (the carrier of N)).n by FUNCT_1:35;
  then n <= (id N).n by GRCAT_1:def 11;
  hence thesis;
end;

theorem Th5:
for N being directed non empty RelStr, x,y being Element of N
  ex z being Element of N st x <= z & y <= z
proof
  let N be directed non empty RelStr, x,y be Element of N;
A1:
  [#]N is directed by WAYBEL_0:def 6;
    x in the carrier of N;
  then A2:  x in [#]N by PRE_TOPC:12;
    y in the carrier of N;
  then y in [#]N by PRE_TOPC:12;
  then ex z being Element of N st z in [#]N & x <= z & y <= z
        by A1,A2,WAYBEL_0:def 1;
  hence thesis;
end;

theorem Th6:
for N being directed non empty RelStr holds
  ex p being map of N,N st p is greater_or_equal_to_id
proof
  let N be directed non empty RelStr;
  defpred P[set,set] means
       for n,m being Element of N st $1=n & $2=m holds n <= m;
A1:
  for e being set st e in the carrier of N
    ex u being set st u in the carrier of N & P[e,u]
  proof
    let e be set;
    assume e in the carrier of N;
    then reconsider e'=e as Element of N;
    consider u' being Element of N such that
A2:    e'<= u' & e' <= u' by Th5;
    take u';
    thus u' in the carrier of N;
    let n,m be Element of N;
    assume
        e=n & u'=m;
    hence n <= m by A2;
  end;
  consider p being Function such that
A3:  dom p = the carrier of N and
A4:  rng p c= the carrier of N and
A5:  for e being set st e in the carrier of N holds P[e,p.e]
         from NonUniqBoundFuncEx(A1);
    p is Function of the carrier of N, the carrier of N
    by A3,A4,FUNCT_2:4;
  then reconsider p as map of N,N;
  take p;
  let j be Element of N;
  reconsider j'=j as Element of N;
    for n,m being Element of N st j'=n & p.j'=m holds n <= m by A5;
  hence j <= p.j;
end;

definition
  let N be directed non empty RelStr;
  cluster greater_or_equal_to_id map of N,N;
  existence by Th6;
end;

definition
  let N be reflexive non empty RelStr;
  cluster greater_or_equal_to_id map of N,N;
existence
proof
  take id N;
  thus id N is greater_or_equal_to_id by Th4;
end;
end;

definition
  let L be non empty 1-sorted;
  let N be non empty NetStr over L;
  let f be map of N,N;
  func N * f -> strict non empty NetStr over L means :Def2:
    the RelStr of it = the RelStr of N &
    the mapping of it = (the mapping of N) * f;
existence
proof
 take NetStr (# the carrier of N,the InternalRel of N,(the mapping of N)*f #);
 thus thesis;
end;
uniqueness;
end;

theorem Th7:
for L being non empty 1-sorted, N being non empty NetStr over L,
    f being map of N, N holds
  the carrier of N * f = the carrier of N
proof
  let L be non empty 1-sorted, N be non empty NetStr over L, f be map of N, N;
    the RelStr of N * f = the RelStr of N by Def2;
  hence the carrier of N * f = the carrier of N;
end;

theorem Th8:
for L being non empty 1-sorted, N being non empty NetStr over L,
    f being map of N,N holds
  N * f = NetStr(#the carrier of N,the InternalRel of N,(the mapping of N)*f#)
proof
  let L be non empty 1-sorted, N be non empty NetStr over L,
      f be map of N,N;
    the RelStr of N*f=the RelStr of N by Def2;
  hence
    N * f=NetStr(#the carrier of N,the InternalRel of N,(the mapping of N)*f#)
      by Def2;
end;

theorem Th9:
for L being non empty 1-sorted,
    N being transitive directed non empty RelStr,
    f being Function of the carrier of N,the carrier of L
holds NetStr (#the carrier of N,the InternalRel of N,f#) is net of L
proof
  let L be non empty 1-sorted,
      N be transitive directed non empty RelStr,
      f be Function of the carrier of N,the carrier of L;
  set N2 = NetStr(#the carrier of N,the InternalRel of N,f#);
A1: the RelStr of N = the RelStr of N2;
A2: [#]N = the carrier of N by PRE_TOPC:12
    .= [#]N2 by PRE_TOPC:12;
    [#]N is directed by WAYBEL_0:def 6;
  then [#]N2 is directed by A1,A2,WAYBEL_0:3;
  hence NetStr (#the carrier of N,the InternalRel of N,f#) is net of L
     by A1,WAYBEL_0:def 6,WAYBEL_8:13;
end;

definition
let L be non empty 1-sorted,
    N be transitive directed non empty RelStr,
    f be Function of the carrier of N,the carrier of L;
  cluster NetStr (#the carrier of N,the InternalRel of N,f#) ->
    transitive directed non empty;
correctness by Th9;
end;

theorem Th10:
for L being non empty 1-sorted, N being net of L, p being map of N,N holds
  N * p is net of L
proof
  let L be non empty 1-sorted, N be net of L, p be map of N,N;
    N * p = NetStr(#the carrier of N,the InternalRel of N,(the mapping of N)*p
#)
    by Th8;
  hence N * p is net of L;
end;

definition
  let L be non empty 1-sorted, N be net of L;
  let p be map of N,N;
  cluster N * p -> transitive directed;
  correctness by Th10;
end;

theorem Th11:
for L being non empty 1-sorted, N being net of L, p being map of N,N
st N in NetUniv L holds
  N * p in NetUniv L
proof
  let L be non empty 1-sorted, N be net of L, p be map of N,N;
  assume N in NetUniv L;
  then consider N1 being strict net of L such that
A1:  N1 = N and
A2:  the carrier of N1 in the_universe_of the carrier of L by YELLOW_6:def 14;
    the carrier of N * p = the carrier of N by Th7;
  hence N * p in NetUniv L by A1,A2,YELLOW_6:def 14;
end;

theorem Th12:
for L being non empty 1-sorted, N,M being net of L
  st the NetStr of N = the NetStr of M holds
    M is subnet of N
proof
  let L be non empty 1-sorted, N,M be net of L;
  assume
A1: the NetStr of N = the NetStr of M;
A2:
  N is subnet of N by YELLOW_6:23;
    ex f being map of M, N st
    the mapping of M = (the mapping of N)*f &
    for m being Element of N ex n being Element of M st
      for p being Element of M st n <= p holds m <= f.p
  proof
    consider f being map of N, N such that
A3:   the mapping of N = (the mapping of N)*f and
A4:    for m being Element of N ex n being Element of N st
        for p being Element of N st n <= p holds m <= f.p
          by A2,YELLOW_6:def 12;
    reconsider f2=f as map of M,N by A1;
    take f2;
    thus the mapping of M = (the mapping of N)*f2 by A1,A3;
    let m be Element of N;
    consider n being Element of N such that
A5:    for p being Element of N st n <= p holds m <= f.p by A4;
    reconsider n2=n as Element of M by A1;
    take n2;
    let p be Element of M;
    reconsider p1=p as Element of N by A1;
    assume n2 <= p;
    then [n2,p] in the InternalRel of M by ORDERS_1:def 9;
    then n <= p1 by A1,ORDERS_1:def 9;
    hence m <= f2.p by A5;
  end;
  hence M is subnet of N by YELLOW_6:def 12;
end;

definition
  let L be non empty 1-sorted, N be net of L;
  cluster strict subnet of N;
existence
proof
  reconsider
    N2= NetStr (#the carrier of N,the InternalRel of N,the mapping of N#)
      as subnet of N by Th12;
  take N2;
  thus N2 is strict;
end;
end;

theorem Th13:
for L being non empty 1-sorted, N being net of L,
    p being greater_or_equal_to_id map of N,N holds N * p is subnet of N
proof
  let L be non empty 1-sorted;
  let N be net of L;
  let p be greater_or_equal_to_id map of N,N;
    ex f being map of (N * p), N st
   the mapping of (N * p) = (the mapping of N)*f &
   for m being Element of N ex n being Element of (N * p) st
    for k being Element of (N * p) st n <= k holds m <= f.k
  proof
      the carrier of N * p = the carrier of N by Th7;
    then reconsider f=p as map of (N * p), N;
    take f;
    thus the mapping of (N * p) = (the mapping of N)*f by Def2;
    let m be Element of N;
      m in the carrier of N;
    then m in the carrier of N*p by Th7;
    then reconsider n=m as Element of (N * p);
    take n;
    let k be Element of (N * p);
    assume
A1:    n <= k;
      k in the carrier of N*p;
    then k in the carrier of N by Th7;
    then reconsider k1 = k as Element of N;
      the RelStr of N*p = the RelStr of N by Def2;
    then A2:    m <= k1 by A1,YELLOW_0:1;
      k1 <= p.k1 by Def1;
    hence m <= f.k by A2,YELLOW_0:def 2;
  end;
  hence N * p is subnet of N by YELLOW_6:def 12;
end;

definition
::3.1 Proposition (2)=>(3)
  let L be non empty 1-sorted;
  let N be net of L;
  let p be greater_or_equal_to_id map of N,N;
  redefine func N * p -> strict subnet of N;
  correctness by Th13;
end;

theorem ::3.1 Proposition (2b)=>(3)
Th14:
for L being complete LATTICE, N being net of L, x being Element of L st
  N in NetUniv L
holds
    (x=lim_inf N &
     for M being subnet of N st M in NetUniv L holds x >= inf M)
implies
      x=lim_inf N &
      for p being greater_or_equal_to_id map of N,N holds x >= inf (N * p)
proof
  let L be complete LATTICE, N be net of L, x be Element of L;
  assume
A1:  N in NetUniv L;
  assume that
A2: x=lim_inf N and
A3: for M being subnet of N st M in NetUniv L
      holds x >= inf M;
  thus x=lim_inf N by A2;
  let p be greater_or_equal_to_id map of N,N;
    N * p in NetUniv L by A1,Th11;
  hence x >= inf (N * p) by A3;
end;

theorem ::3.1 Proposition (3)=>(1)
Th15:
for L being complete LATTICE, N being net of L, x being Element of L holds
  (x=lim_inf N &
   for p being greater_or_equal_to_id map of N,N holds x >= inf (N * p) )
     implies for M being subnet of N holds x = lim_inf M
proof
  let L be complete LATTICE, N be net of L, x be Element of L;
  assume that
A1: x=lim_inf N and
A2: for p being greater_or_equal_to_id map of N,N holds x >= inf (N * p);
  let M be subnet of N;
  consider f being map of M, N such that
A3: the mapping of M = (the mapping of N)*f and
A4: for j being Element of N ex k being Element of M st
      for m being Element of M st k <= m holds j <= f.m by YELLOW_6:def 12;
A5:
  x <= lim_inf M by A1,WAYBEL21:37;
A6:
  for k0 being Element of M holds
    "/\"({M.k where k is Element of M:k >= k0},L)<=x
  proof
    let k0 be Element of M;
A7:  for j being Element of N ex v being Element of M
      st v >= k0 & for v' being Element of M st v'>= v
        holds f.(v')>=j & f.v >= j
    proof
      let j be Element of N;
      consider w being Element of M such that
A8:      for w' being Element of M st w <= w' holds j <= f.(w') by A4;
      consider v being Element of M such that
A9:     v >= k0 and
A10:     v >= w by Th5;
      take v;
      thus v >= k0 by A9;
      let v' be Element of M;
      assume v'>= v;
      then v' >= w by A10,YELLOW_0:def 2;
      hence f.(v')>=j by A8;
      thus f.v >= j by A8,A10;
    end;
    defpred P[set,set] means
       for j being Element of N, v,v' being Element of M
             st $1=j & $2=v & v'>= v holds v >= k0 & f.(v')>=j & f.v >= j;
A11:  for e being set st e in the carrier of N
      ex u being set st u in the carrier of M & P[e,u]
    proof
      let e be set;
      assume e in the carrier of N;
      then reconsider e'=e as Element of N;
      consider u being Element of M such that
A12:     u >= k0 and
A13:     for v' being Element of M st v'>= u holds f.(v')>=e' & f.u >= e' by A7
;
      take u;
      thus u in the carrier of M;
      let j be Element of N, v,v' be Element of M;
      assume that
A14:     e=j and
A15:     u=v and
A16:     v'>= v;
      thus v >= k0 by A12,A15;
      thus f.(v')>=j by A13,A14,A15,A16;
      thus f.v >= j by A13,A14,A15,A16;
    end;
    consider g being Function such that
A17:   dom g = the carrier of N and
A18:   rng g c= the carrier of M and
A19:   for e being set st e in the carrier of N holds P[e,g.e]
           from NonUniqBoundFuncEx(A11);
    reconsider g as Function of the carrier of N,the carrier of M by A17,A18,
FUNCT_2:4;
A20:
    for j being Element of N holds g.j >= k0
    proof
      let j be Element of N;
      reconsider v=g.j as Element of M;
        ex v' being Element of M st v'>= v & v'>= v by Th5;
      then consider v' being Element of M such that
A21:    v'>= v;
      thus g.j >= k0 by A19,A21;
    end;
    reconsider g as map of N,M;
    reconsider p=f*g as map of N,N;
      for j being Element of N holds j <= p.j
    proof
      let j be Element of N;
      reconsider j1=j as Element of N;
A22:   for v,v' being Element of M
         st j=j1 & g.j=v & v'>= v holds v >= k0 & f.(v')>=j & f.v >= j by A19;
      reconsider v=g.j as Element of M;
A23:  [#]M is directed by WAYBEL_0:def 6;
        v in the carrier of M;
      then v in [#]M by PRE_TOPC:12;
      then consider v' being Element of M such that
A24:     v' in [#]M & v <= v' & v <=v' by A23,WAYBEL_0:def 1;
        j <= f.(g.j) by A22,A24;
      hence j <= p.j by A17,FUNCT_1:23;
    end;
    then reconsider p as greater_or_equal_to_id map of N,N by Def1;
A25: ex_inf_of {(N*p).j where j is Element of (N*p):
                 not contradiction},L by YELLOW_0:17;
A26: ex_inf_of {M.k where k is Element of M:k >= k0},L
          by YELLOW_0:17;
A27:
    {(N*p).j where j is Element of (N*p):not contradiction} c=
      {M.k where k is Element of M:k >= k0}
    proof
      let y be set;
      assume
          y in {(N*p).j where j is Element of (N*p):
         not contradiction};
      then consider j being Element of (N*p) such that
A28:    y = (N*p).j;
A29:  the carrier of (N*p)= the carrier of N by Th7;
A30:  y = (the mapping of (N*p)).j by A28,WAYBEL_0:def 8
       .= ((the mapping of N)*(f*g)).j by Def2
       .= ((the mapping of M)*g).j by A3,RELAT_1:55
       .= (the mapping of M).(g.j) by A17,A29,FUNCT_1:23;
      reconsider j'=j as Element of N by A29;
A31:  y = M.(g.j') by A30,WAYBEL_0:def 8;
        g.j' >= k0 by A20;
      hence y in {M.k where k is Element of M:k >= k0} by A31;
    end;
A32:
    dom (the mapping of (N*p)) = the carrier of (N*p) by FUNCT_2:def 1;
A33: rng the mapping of (N*p) =
      {(N*p).j where j is Element of (N*p): not contradiction}
    proof
      thus rng the mapping of (N*p) c=
        {(N*p).j where j is Element of (N*p):
         not contradiction}
      proof
        let y be set;
        assume y in rng the mapping of (N*p);
        then consider j1 being set such that
A34:      j1 in dom (the mapping of (N*p)) and
A35:      (the mapping of (N*p)).j1 = y by FUNCT_1:def 5;
        reconsider j1 as Element of (N*p) by A34;
          y = (N*p).j1 by A35,WAYBEL_0:def 8;
        hence y in {(N*p).j where j is Element of (N*p):
         not contradiction};
      end;
      let y be set;
      assume
          y in {(N*p).j where j is Element of (N*p):
        not contradiction};
      then consider j being Element of (N*p) such that
A36:    y = (N*p).j;
        y = (the mapping of (N*p)).j by A36,WAYBEL_0:def 8;
      hence y in rng the mapping of (N*p) by A32,FUNCT_1:def 5;
    end;
      inf (N*p) = Inf the mapping of (N*p) by WAYBEL_9:def 2
             .= "/\"({(N*p).j where j is Element of (N*p):
                     not contradiction},L) by A33,YELLOW_2:def 6;
    then A37:   "/\"({M.k where k is Element of M:k >= k0},L)<=
inf (N*p)
             by A25,A26,A27,YELLOW_0:35;
      inf (N * p) <= x by A2;
    hence "/\"({M.k where k is Element of M:k >= k0},L)<=x
      by A37,YELLOW_0:def 2;
  end;
    for b being Element of L
    st b in { "/\"({M.k where k is Element of M:k >= k0},L)
            where k0 is Element of M: not contradiction}
  holds b <= x
  proof
    let b be Element of L;
    assume b in { "/\"({M.k where k is Element of M:k >= k0},L)
                where k0 is Element of M: not contradiction};
    then consider k0 being Element of M such that
A38:    b = "/\"({M.k where k is Element of M:k >= k0},L);
    reconsider k0 as Element of M;
      "/\"({M.k where k is Element of M:k >= k0},L)<=x by A6;
    hence b <= x by A38;
  end;
  then A39:  x is_>=_than { "/\"
({M.k where k is Element of M:k >= k0},L)
                 where k0 is Element of M: not contradiction}
                   by LATTICE3:def 9;
    ex_sup_of
   { "/\"({M.k where k is Element of M:k >= k0},L)
    where k0 is Element of M: not contradiction},L
      by YELLOW_0:17;
  then "\/"({ "/\"({M.k where k is Element of M:k >= k0},L)
           where k0 is Element of M: not contradiction},L)<=x
             by A39,YELLOW_0:def 9;
  then x >= lim_inf M by WAYBEL11:def 6;
  hence x = lim_inf M by A5,ORDERS_1:25;
end;

definition
let L be non empty RelStr;
func lim_inf-Convergence L -> Convergence-Class of L means :Def3:
 for N being net of L st N in NetUniv L
 for x being Element of L
  holds [N,x] in it iff for M being subnet of N holds x = lim_inf M;
existence
proof
  defpred P[set,set] means
    ex N being strict net of L st
      $1=N & for M being subnet of N holds $2 = lim_inf M;
  consider C being Relation of NetUniv L,the carrier of L such that
A1:  for N' being Element of NetUniv L, x being Element of L
    holds [N',x] in C iff P[N', x] from Rel_On_Dom_Ex;
  reconsider C as Convergence-Class of L by YELLOW_6:def 21;
  take C;
  let N be net of L;
  assume N in NetUniv L;
  then reconsider N1=N as Element of NetUniv L;
  let x be Element of L;
  thus [N,x] in C implies for M being subnet of N holds x = lim_inf M
  proof
    assume
A2:    [N,x] in C;
    let M be subnet of N;
    consider N2 being strict net of L such that
A3:   N1=N2 and
A4:   for M being subnet of N2 holds x = lim_inf M by A1,A2;
    thus x = lim_inf M by A3,A4;
  end;
  assume
A5:  for M being subnet of N holds x = lim_inf M;
  consider N2 being strict net of L such that
A6:  N2=N1 and
      the carrier of N2 in the_universe_of the carrier of L by YELLOW_6:def 14;
  thus [N,x] in C by A1,A5,A6;
end;
uniqueness
proof
  let C1,C2 be Convergence-Class of L such that
A7:  for N being net of L st N in NetUniv L
       for x being Element of L
       holds [N,x] in C1 iff for M being subnet of N holds x = lim_inf M and
A8:  for N being net of L st N in NetUniv L
       for x being Element of L
       holds [N,x] in C2 iff for M being subnet of N holds x = lim_inf M;
A9:
  C1 c= [:NetUniv L,the carrier of L:] by YELLOW_6:def 21;
A10:
  C2 c= [:NetUniv L,the carrier of L:] by YELLOW_6:def 21;
  let Ns,xs be set;
  thus [Ns,xs] in C1 implies [Ns,xs] in C2
  proof
    assume
A11:   [Ns,xs] in C1;
    then Ns in NetUniv L by A9,ZFMISC_1:106;
    then consider N being strict net of L such that
A12:    N = Ns and
        the carrier of N in the_universe_of the carrier of L
               by YELLOW_6:def 14;
A13: N in NetUniv L by A9,A11,A12,ZFMISC_1:106;
    reconsider x=xs as Element of L by A9,A11,ZFMISC_1:106;
      for M being subnet of N holds x = lim_inf M by A7,A11,A12,A13;
    hence [Ns,xs] in C2 by A8,A12,A13;
  end;
  assume
A14: [Ns,xs] in C2;
  then Ns in NetUniv L by A10,ZFMISC_1:106;
  then consider N being strict net of L such that
A15:  N = Ns and
      the carrier of N in the_universe_of the carrier of L by YELLOW_6:def 14;
A16:
  N in NetUniv L by A10,A14,A15,ZFMISC_1:106;
  reconsider x=xs as Element of L by A10,A14,ZFMISC_1:106;
    for M being subnet of N holds x = lim_inf M by A8,A14,A15,A16;
  hence [Ns,xs] in C1 by A7,A15,A16;
end;
end;

theorem
  for L being complete LATTICE,
    N being net of L,
    x being Element of L
st N in NetUniv L
holds [N,x] in lim_inf-Convergence L iff
  for M being subnet of N st M in NetUniv L holds x = lim_inf M
proof
  let L be complete LATTICE;
  let N be net of L;
  let x be Element of L;
  assume
A1:  N in NetUniv L;
  hence [N,x] in lim_inf-Convergence L implies
     for M being subnet of N st M in NetUniv L holds x = lim_inf M by Def3;
  assume for M being subnet of N st M in NetUniv L holds x = lim_inf M;
  then x=lim_inf N &
       for M being subnet of N st M in NetUniv L holds x >= inf M
           by A1,Th3;
  then x=lim_inf N &
      for p being greater_or_equal_to_id map of N,N holds x >= inf (N * p)
         by A1,Th14;
  then for M being subnet of N holds x = lim_inf M by Th15;
  hence [N,x] in lim_inf-Convergence L by A1,Def3;
end;

theorem Th17:
for L being non empty RelStr,
    N being constant net of L,
    M being subnet of N
holds M is constant & the_value_of N = the_value_of M
proof
  let L be non empty RelStr,
      N be constant net of L,
      M be subnet of N;
  consider f being map of M, N such that
A1: the mapping of M = (the mapping of N)*f and
      for m being Element of N ex n being Element of M st
      for p being Element of M st n <= p holds m <= f.p
          by YELLOW_6:def 12;
A2:
  dom (the mapping of M)=the carrier of M by FUNCT_2:def 1;
A3:
  dom f = the carrier of M by FUNCT_2:def 1;
A4:
  dom (the mapping of N) = the carrier of N by FUNCT_2:def 1;
    for n1,n2 being set st n1 in dom (the mapping of M) &
                         n2 in dom (the mapping of M) holds
     (the mapping of M).n1=(the mapping of M).n2
  proof
    let n1,n2 be set;
    assume that
A5:   n1 in dom (the mapping of M) and
A6:   n2 in dom (the mapping of M);
A7:
    f.n1 in rng f by A3,A5,FUNCT_1:def 5;
A8:
    f.n2 in rng f by A3,A6,FUNCT_1:def 5;
    thus (the mapping of M).n1= (the mapping of N).(f.n1) by A1,A3,A5,FUNCT_1:
23
      .= (the mapping of N).(f.n2) by A4,A7,A8,SEQM_3:def 5
      .= (the mapping of M).n2 by A1,A3,A6,FUNCT_1:23;
  end;
  then A9: the mapping of M is constant by SEQM_3:def 5;
  hence
A10:  M is constant by YELLOW_6:def 6;
  consider x being set such that
A11: x in dom (the mapping of N) and
A12: the_value_of the mapping of N = (the mapping of N).x by YELLOW_6:def 1;
  consider y being Element of dom (the mapping of M);

   f.y in rng f by A2,A3,FUNCT_1:def 5;
then A13:
  the_value_of the mapping of N= (the mapping of N).(f.y) by A4,A11,A12,SEQM_3:
def 5
          .= (the mapping of M).y by A1,A2,FUNCT_1:22;
  thus the_value_of N = the_value_of the mapping of N by YELLOW_6:def 10
         .= the_value_of the mapping of M by A2,A9,A13,YELLOW_6:def 1
         .= the_value_of M by A10,YELLOW_6:def 10;
end;

definition let L be non empty RelStr;
func xi L -> Subset-Family of L equals :Def4:
  the topology of ConvergenceSpace lim_inf-Convergence L;
correctness by YELLOW_6:def 27;
end;

theorem
  for L being complete LATTICE
holds lim_inf-Convergence L is (CONSTANTS)
proof
  let L be complete LATTICE;
  let N be constant net of L;
  assume
A1:  N in NetUniv L;
    for M being subnet of N holds the_value_of N = lim_inf M
  proof
    let M be subnet of N;
A2:  M is constant by Th17;
    thus the_value_of N = the_value_of M by Th17
                        .= lim_inf M by A2,WAYBEL11:23;
  end;
  hence [N,the_value_of N] in lim_inf-Convergence L by A1,Def3;
end;

theorem
  for L being non empty RelStr holds
  lim_inf-Convergence L is (SUBNETS)
proof
  let L be non empty RelStr;
  let N be net of L, M be subnet of N;
  assume
A1:  M in NetUniv L;
  let x be Element of L;
  assume
A2:  [N,x] in lim_inf-Convergence L;
    lim_inf-Convergence L c= [:NetUniv L,the carrier of L:] by YELLOW_6:def 21;
  then A3:  N in NetUniv L by A2,ZFMISC_1:106;
    for M1 being subnet of M holds x = lim_inf M1
  proof
    let M1 be subnet of M;
    reconsider M1'=M1 as subnet of N by YELLOW_6:24;
      x = lim_inf M1' by A2,A3,Def3;
    hence x = lim_inf M1;
  end;
  hence [M,x] in lim_inf-Convergence L by A1,Def3;
end;

theorem
  for L being continuous complete LATTICE holds
  lim_inf-Convergence L is (DIVERGENCE)
proof
  let L be continuous complete LATTICE;
  let N be net of L, x be Element of L;
  assume
A1:  N in NetUniv L & not [N,x] in lim_inf-Convergence L;
  then not for M being subnet of N holds x = lim_inf M by Def3;
  then A2:
    not (x=lim_inf N &
    for p being greater_or_equal_to_id map of N,N holds x >= inf (N * p))
      by Th15;
A3:
  lim_inf-Convergence L c= [:NetUniv L,the carrier of L:] by YELLOW_6:def 21;
  consider N1 being strict net of L such that
A4: N1=N and
      the carrier of N1 in
 the_universe_of the carrier of L by A1,YELLOW_6:def 14;
  per cases by A1,A2,Th14;
  suppose
A5: not (x=lim_inf N) & x<=lim_inf N;
  reconsider N'=N as subnet of N by YELLOW_6:23;
  take N';
  thus N' in NetUniv L by A1;
  given M2 being subnet of N' such that
A6:  [M2,x] in lim_inf-Convergence L;
A7:
  M2 in NetUniv L by A3,A6,ZFMISC_1:106;
    M2 is subnet of M2 by YELLOW_6:23;
  then A8: lim_inf M2 =x by A6,A7,Def3;
    lim_inf N <= lim_inf M2 by WAYBEL21:37;
  hence contradiction by A5,A8,YELLOW_0:def 3;
  suppose not (x=lim_inf N) & not (x<=lim_inf N);
  then not x is_S-limit_of N by WAYBEL11:def 7;
  then not [N,x] in Scott-Convergence L by A1,A4,WAYBEL11:def 8;
  then consider M being subnet of N such that
A9:  M in NetUniv L and
A10:  not ex M1 being subnet of M st [M1,x] in Scott-Convergence L
               by A1,YELLOW_6:def 25;
  take M;
  thus M in NetUniv L by A9;
    for M1 being subnet of M holds not [M1,x] in lim_inf-Convergence L
  proof
    let M1 be subnet of M;
    assume
A11:   [M1,x] in lim_inf-Convergence L;
    then A12:   M1 in NetUniv L by A3,ZFMISC_1:106;
    then consider M11 being strict net of L such that
A13:   M11=M1 and
        the carrier of M11 in
 the_universe_of the carrier of L by YELLOW_6:def 14;
      M1 is subnet of M1 by YELLOW_6:23;
    then A14: x = lim_inf M1 by A11,A12,Def3;
      not [M1,x] in Scott-Convergence L by A10;
    then not x is_S-limit_of M1 by A12,A13,WAYBEL11:def 8;
    hence contradiction by A14,WAYBEL11:def 7;
  end;
  hence not ex M1 being subnet of M st [M1,x] in lim_inf-Convergence L;
  suppose not (for M being subnet of N st M in NetUniv L holds x >= inf M);
  then consider M being subnet of N such that
A15: M in NetUniv L and
A16:  not x >= inf M;
  take M;
  thus M in NetUniv L by A15;
    for M1 being subnet of M holds not [M1,x] in lim_inf-Convergence L
  proof
    let M1 be subnet of M;
    assume
A17:    [M1,x] in lim_inf-Convergence L;
    then A18: M1 in NetUniv L by A3,ZFMISC_1:106;
      M1 is subnet of M1 by YELLOW_6:23;
    then A19:   x = lim_inf M1 by A17,A18,Def3;
A20:  lim_inf M1 >= lim_inf M by WAYBEL21:37;
      lim_inf M >= inf M by Th1;
    hence contradiction by A16,A19,A20,YELLOW_0:def 2;
  end;
  hence not ex M1 being subnet of M st [M1,x] in lim_inf-Convergence L;
end;

theorem
  for L being non empty RelStr, N,x being set holds
  [N,x] in lim_inf-Convergence L implies N in NetUniv L
proof
  let L be non empty RelStr, N,x be set;
  assume
A1:  [N,x] in lim_inf-Convergence L;
    lim_inf-Convergence L c= [:NetUniv L,the carrier of L:] by YELLOW_6:def 21;
  hence N in NetUniv L by A1,ZFMISC_1:106;
end;

theorem Th22:
for L being non empty 1-sorted, C1,C2 being Convergence-Class of L holds
  C1 c= C2 implies
   the topology of ConvergenceSpace C2 c= the topology of ConvergenceSpace C1
proof
  let L be non empty 1-sorted, C1,C2 be Convergence-Class of L;
  assume
A1: C1 c= C2;
  let A be set;
  assume A in the topology of ConvergenceSpace C2;
  then A in { V where V is Subset of L:
   for p being Element of L st p in V
    for N being net of L st [N,p] in C2 holds N is_eventually_in V}
      by YELLOW_6:def 27;
  then consider V1 being Subset of L such that
A2: A=V1 and
A3: for p being Element of L st p in V1
      for N being net of L st [N,p] in C2 holds N is_eventually_in V1;
    ex V being Subset of L st
    A=V &
    for p being Element of L st p in V
      for N being net of L st [N,p] in C1 holds N is_eventually_in V
  proof
    take V1;
    thus A=V1 by A2;
    let p be Element of L;
    assume
A4:   p in V1;
    let N be net of L;
    assume [N,p] in C1;
    hence N is_eventually_in V1 by A1,A3,A4;
  end;
  then A in { V where V is Subset of L:
   for p being Element of L st p in V
    for N being net of L st [N,p] in C1 holds N is_eventually_in V};
  hence A in the topology of ConvergenceSpace C1 by YELLOW_6:def 27;
end;

theorem Th23:
for L being non empty reflexive RelStr holds
  lim_inf-Convergence L c= Scott-Convergence L
proof
  let L be non empty reflexive RelStr;
  let Ns,xs be set;
A1:
  lim_inf-Convergence L c= [:NetUniv L,the carrier of L:] by YELLOW_6:def 21;
  assume
A2:  [Ns,xs] in lim_inf-Convergence L;
  then Ns in NetUniv L by A1,ZFMISC_1:106;
  then consider N being strict net of L such that
A3:  N = Ns and
      the carrier of N in the_universe_of the carrier of L by YELLOW_6:def 14;
A4:
  N in NetUniv L by A1,A2,A3,ZFMISC_1:106;
  reconsider x=xs as Element of L by A1,A2,ZFMISC_1:106;
    N is subnet of N by YELLOW_6:23;
  then x = lim_inf N by A2,A3,A4,Def3;
  then x is_S-limit_of N by WAYBEL11:def 7;
  hence [Ns,xs] in Scott-Convergence L by A3,A4,WAYBEL11:def 8;
end;

theorem Th24:
for X,Y being set holds
  X c= Y implies X in the_universe_of Y
proof
  let X,Y be set;
  assume
A1:  X c= Y;
    Tarski-Class the_transitive-closure_of Y
    is_Tarski-Class_of the_transitive-closure_of Y by CLASSES1:def 4;
  then A2:  the_transitive-closure_of Y in Tarski-Class
the_transitive-closure_of Y
      by CLASSES1:def 3;
    Y c= the_transitive-closure_of Y by CLASSES1:59;
  then X c= the_transitive-closure_of Y by A1,XBOOLE_1:1;
  then X in Tarski-Class the_transitive-closure_of Y by A2,CLASSES1:def 1;
  hence X in the_universe_of Y by YELLOW_6:def 3;
end;

theorem Th25:
for L being non empty transitive reflexive RelStr,
    D being directed non empty Subset of L holds
Net-Str D in NetUniv L
proof
  let L be non empty transitive reflexive RelStr;
  let D be directed non empty Subset of L;
A1:
  D in the_universe_of the carrier of L by Th24;
    the carrier of Net-Str D = D by WAYBEL21:32;
  hence Net-Str D in NetUniv L by A1,YELLOW_6:def 14;
end;


theorem Th26:
for L being complete LATTICE,
    D being directed non empty Subset of L holds
for M being subnet of Net-Str D holds
        lim_inf M = sup D
proof
  let L be complete LATTICE;
  let D be directed non empty Subset of L;

   for M being subnet of Net-Str D holds sup D >= inf M
  proof
    let M be subnet of Net-Str D;
    consider i being Element of M;
    set f=the mapping of M;
    consider g being map of M, Net-Str D such that
A1:   the mapping of M = (the mapping of Net-Str D)*g and
        for m being Element of Net-Str D ex n being Element of M st
        for p being Element of M st n <= p holds m <= g.p by YELLOW_6:def 12;
A2:
    dom f = the carrier of M by FUNCT_2:def 1;
      g.i in the carrier of Net-Str D;
    then A3:    g.i in D by WAYBEL21:32;
    then g.i = (id D).(g.i) by FUNCT_1:35
            .= (the mapping of Net-Str D).(g.i) by WAYBEL21:32
            .= f.i by A1,A2,FUNCT_1:22;
    then A4:    f.i <= sup D by A3,YELLOW_2:24;
A5: ex_inf_of rng f,L by YELLOW_0:17;
     f.i in rng f by A2,FUNCT_1:def 5;
    then "/\"(rng f,L) <= f.i by A5,YELLOW_4:2;
    then sup D >= "/\"(rng f,L) by A4,YELLOW_0:def 2;
    then sup D >= Inf the mapping of M by YELLOW_2:def 6;
    hence sup D >= inf M by WAYBEL_9:def 2;
  end;
  then lim_inf Net-Str D =sup D &
   for p being greater_or_equal_to_id map of Net-Str(D),Net-Str(D) holds
        sup D >= inf (Net-Str(D)*p) by WAYBEL17:10;
  hence for M being subnet of Net-Str D holds
        lim_inf M = sup D by Th15;
end;

theorem Th27:
for L being non empty complete LATTICE,
    D being directed non empty Subset of L holds
[Net-Str D,sup D] in lim_inf-Convergence L
proof
  let L be non empty complete LATTICE;
  let D be directed non empty Subset of L;
A1:
  Net-Str D in NetUniv L by Th25;
    for M being subnet of Net-Str D holds lim_inf M = sup D by Th26;
  hence [Net-Str D,sup D] in lim_inf-Convergence L by A1,Def3;
end;

theorem Th28:
for L being complete LATTICE, U1 being Subset of L holds
U1 in xi(L) implies U1 is property(S)
proof
  let L be complete LATTICE;
  let U1 be Subset of L;
  assume U1 in xi(L);
  then U1 in the topology of ConvergenceSpace lim_inf-Convergence L by Def4;
  then U1 in { V where V is Subset of L:
    for p being Element of L st p in V
    for N being net of L st [N,p] in lim_inf-Convergence L holds
         N is_eventually_in V} by YELLOW_6:def 27;
  then consider V being Subset of L such that
A1: U1=V and
A2: for p being Element of L st p in V
    for N being net of L st [N,p] in lim_inf-Convergence L holds
         N is_eventually_in V;
  let D be non empty directed Subset of L;
  assume
A3:  sup D in U1;
    [Net-Str D,sup D] in lim_inf-Convergence L by Th27;
  then (Net-Str D) is_eventually_in U1 by A1,A2,A3;
  then consider y being Element of (Net-Str D) such that
A4:  for x being Element of (Net-Str D) st y <= x holds
      (Net-Str D).x in U1 by WAYBEL_0:def 11;
A5:
  y in the carrier of Net-Str D;
  then y in D by WAYBEL21:32;
  then reconsider y1=y as Element of L;
  reconsider y1 as Element of L;
  take y1;
  thus y1 in D by A5,WAYBEL21:32;
  let x1 be Element of L;
  assume that
A6: x1 in D and
A7: x1 >= y1;
  reconsider x=x1 as Element of Net-Str D by A6,WAYBEL21:32;
  reconsider x as Element of Net-Str D;
    Net-Str D is full SubRelStr of L by WAYBEL21:32;
  then A8:  y <= x by A7,YELLOW_0:61;
    (Net-Str D).x = (the mapping of Net-Str D).x by WAYBEL_0:def 8
               .= (id D).x by WAYBEL21:32
               .= x by A6,FUNCT_1:35;
  hence x1 in U1 by A4,A8;
end;

theorem Th29:
for L being non empty reflexive RelStr, A being Subset of L holds
  A in sigma L implies A in xi L
proof
  let L be non empty reflexive RelStr, A be Subset of L;
    lim_inf-Convergence L c= Scott-Convergence L by Th23;
  then A1:  the topology of ConvergenceSpace Scott-Convergence L c=
    the topology of ConvergenceSpace lim_inf-Convergence L by Th22;
  assume A in sigma L;
  then A in the topology of ConvergenceSpace Scott-Convergence L
           by WAYBEL11:def 12;
  then A in the topology of ConvergenceSpace lim_inf-Convergence L
    by A1;
  hence A in xi L by Def4;
end;

theorem Th30:
for L being complete LATTICE, A being Subset of L
st A is upper holds A in xi L implies A in sigma L
proof
  let L be complete LATTICE, A be Subset of L;
  assume
A1: A is upper;
  assume A in xi L;
  then A2: A is property(S) by Th28;
  consider T being Scott TopAugmentation of L;
A3:
  the RelStr of T = the RelStr of L by YELLOW_9:def 4;
  then reconsider A'=A as Subset of T;
  reconsider A' as Subset of T;
A4:
  A' is property(S) by A2,A3,YELLOW12:19;

    A' is upper by A1,A3,WAYBEL_0:25;
  then A' is open by A4,WAYBEL11:15;
  then A' in the topology of T by PRE_TOPC:def 5;
  hence A in sigma L by YELLOW_9:51;
end;

theorem ::3.3 Proposition (ii)
  for L being complete LATTICE , A being Subset of L
st A is lower holds A` in xi L iff A is closed_under_directed_sups
proof
  let L be complete LATTICE, A be Subset of L;
  assume
A1:  A is lower;
  then reconsider A1=A as lower Subset of L;
A2:
  A1` is upper;
  consider T being Scott TopAugmentation of L;
A3:
  the RelStr of L = the RelStr of T by YELLOW_9:def 4;
  then reconsider A'=A as Subset of T;
  reconsider A' as Subset of T;

   [#]L = the carrier of T by A3,PRE_TOPC:12
    .= [#]T by PRE_TOPC:12;
then A4:
  A` = [#]T \ A' by PRE_TOPC:17
    .= A'` by PRE_TOPC:17;
  thus A` in xi L implies A is closed_under_directed_sups
  proof
    assume
      A` in xi L;
    then A'` in sigma L by A2,A4,Th30;
    then A'` in the topology of T by YELLOW_9:51;
    then A'` is open by PRE_TOPC:def 5;
    then A' is closed by TOPS_1:29;
    then A' is directly_closed by WAYBEL11:7;
    hence A is closed_under_directed_sups by A3,YELLOW12:20;
  end;
  assume
    A is closed_under_directed_sups;
then A5:
  A' is directly_closed by A3,YELLOW12:20;

    A' is lower by A1,A3,WAYBEL_0:25;
  then A' is closed by A5,WAYBEL11:7;
  then A'` is open by TOPS_1:29;
  then A'` in the topology of T by PRE_TOPC:def 5;
  then A` in sigma L by A4,YELLOW_9:51;
  hence A` in xi L by Th29;
end;

Back to top