Theory Follows

(*  Title:      ZF/UNITY/Follows.thy
    Author:     Sidi O Ehmety, Cambridge University Computer Laboratory
    Copyright   2002  University of Cambridge

Theory ported from HOL.
*)

section‹The "Follows" relation of Charpentier and Sivilotte›

theory Follows imports SubstAx Increasing begin

definition
  Follows :: "[i, i, ii, ii]  i"  where
  "Follows(A, r, f, g) 
            Increasing(A, r, g) Int
            Increasing(A, r,f) Int
            Always({s  state. <f(s), g(s)>:r}) Int
           (k  A. {s  state. <k, g(s)>:r} ⟼w {s  state. <k,f(s)>:r})"

abbreviation
  Incr :: "[ii]i" where
  "Incr(f)  Increasing(list(nat), prefix(nat), f)"

abbreviation
  n_Incr :: "[ii]i" where
  "n_Incr(f)  Increasing(nat, Le, f)"

abbreviation
  s_Incr :: "[ii]i" where
  "s_Incr(f)  Increasing(Pow(nat), SetLe(nat), f)"

abbreviation
  m_Incr :: "[ii]i" where
  "m_Incr(f)  Increasing(Mult(nat), MultLe(nat, Le), f)"

abbreviation
  n_Fols :: "[ii, ii]i"   (infixl n'_Fols 65)  where
  "f n_Fols g  Follows(nat, Le, f, g)"

abbreviation
  Follows' :: "[ii, ii, i, i]  i"
        ((_ /Fols _ /Wrt (_ /'/ _)) [60, 0, 0, 60] 60)  where
  "f Fols g Wrt r/A  Follows(A,r,f,g)"


(*Does this hold for "invariant"?*)

lemma Follows_cong:
     "A=A'; r=r'; x. x  state  f(x)=f'(x); x. x  state  g(x)=g'(x)  Follows(A, r, f, g) = Follows(A', r', f', g')"
by (simp add: Increasing_def Follows_def)


lemma subset_Always_comp:
"mono1(A, r, B, s, h); x  state. f(x):A  g(x):A 
   Always({x  state. <f(x), g(x)>  r})<=Always({x  state. <(h comp f)(x), (h comp g)(x)>  s})"
  unfolding mono1_def metacomp_def
apply (auto simp add: Always_eq_includes_reachable)
done

lemma imp_Always_comp:
"F  Always({x  state. <f(x), g(x)>  r});
    mono1(A, r, B, s, h); x  state. f(x):A  g(x):A 
    F  Always({x  state. <(h comp f)(x), (h comp g)(x)>  s})"
by (blast intro: subset_Always_comp [THEN subsetD])


lemma imp_Always_comp2:
"F  Always({x  state. <f1(x), f(x)>  r});
    F  Always({x  state. <g1(x), g(x)>  s});
    mono2(A, r, B, s, C, t, h);
    x  state. f1(x):A  f(x):A  g1(x):B  g(x):B
   F  Always({x  state. <h(f1(x), g1(x)), h(f(x), g(x))>  t})"
apply (auto simp add: Always_eq_includes_reachable mono2_def)
apply (auto dest!: subsetD)
done

(* comp LeadsTo *)

lemma subset_LeadsTo_comp:
"mono1(A, r, B, s, h); refl(A,r); trans[B](s);
        x  state. f(x):A  g(x):A 
  (j  A. {s  state. <j, g(s)>  r} ⟼w {s  state. <j,f(s)>  r}) <=
 (k  B. {x  state. <k, (h comp g)(x)>  s} ⟼w {x  state. <k, (h comp f)(x)>  s})"

apply (unfold mono1_def metacomp_def, clarify)
apply (simp_all (no_asm_use) add: INT_iff)
apply auto
apply (rule single_LeadsTo_I)
prefer 2 apply (blast dest: LeadsTo_type [THEN subsetD], auto)
apply (rotate_tac 5)
apply (drule_tac x = "g (sa) " in bspec)
apply (erule_tac [2] LeadsTo_weaken)
apply (auto simp add: part_order_def refl_def)
apply (rule_tac b = "h (g (sa))" in trans_onD)
apply blast
apply auto
done

lemma imp_LeadsTo_comp:
"F:(j  A. {s  state. <j, g(s)>  r} ⟼w {s  state. <j,f(s)>  r});
    mono1(A, r, B, s, h); refl(A,r); trans[B](s);
    x  state. f(x):A  g(x):A 
   F:(k  B. {x  state. <k, (h comp g)(x)>  s} ⟼w {x  state. <k, (h comp f)(x)>  s})"
apply (rule subset_LeadsTo_comp [THEN subsetD], auto)
done

lemma imp_LeadsTo_comp_right:
"F  Increasing(B, s, g);
  j  A. F: {s  state. <j, f(s)>  r} ⟼w {s  state. <j,f1(s)>  r};
  mono2(A, r, B, s, C, t, h); refl(A, r); refl(B, s); trans[C](t);
  x  state. f1(x):A  f(x):A  g(x):B; k  C 
  F:{x  state. <k, h(f(x), g(x))>  t} ⟼w {x  state. <k, h(f1(x), g(x))>  t}"
  unfolding mono2_def Increasing_def
apply (rule single_LeadsTo_I, auto)
apply (drule_tac x = "g (sa) " and A = B in bspec)
apply auto
apply (drule_tac x = "f (sa) " and P = "λj. F  X(j) ⟼w Y(j)" for X Y in bspec)
apply auto
apply (rule PSP_Stable [THEN LeadsTo_weaken], blast, blast)
apply auto
apply (force simp add: part_order_def refl_def)
apply (force simp add: part_order_def refl_def)
apply (drule_tac x = "f1 (x)" and x1 = "f (sa) " and P2 = "λx y. uB. P (x,y,u)" for P in bspec [THEN bspec])
apply (drule_tac [3] x = "g (x) " and x1 = "g (sa) " and P2 = "λx y. P (x,y)  d (x,y)  t" for P d in bspec [THEN bspec])
apply auto
apply (rule_tac b = "h (f (sa), g (sa))" and A = C in trans_onD)
apply (auto simp add: part_order_def)
done

lemma imp_LeadsTo_comp_left:
"F  Increasing(A, r, f);
  j  B. F: {x  state. <j, g(x)>  s} ⟼w {x  state. <j,g1(x)>  s};
  mono2(A, r, B, s, C, t, h); refl(A,r); refl(B, s); trans[C](t);
  x  state. f(x):A  g1(x):B  g(x):B; k  C 
  F:{x  state. <k, h(f(x), g(x))>  t} ⟼w {x  state. <k, h(f(x), g1(x))>  t}"
  unfolding mono2_def Increasing_def
apply (rule single_LeadsTo_I, auto)
apply (drule_tac x = "f (sa) " and P = "λk. F  Stable (X (k))" for X in bspec)
apply auto
apply (drule_tac x = "g (sa) " in bspec)
apply auto
apply (rule PSP_Stable [THEN LeadsTo_weaken], blast, blast)
apply auto
apply (force simp add: part_order_def refl_def)
apply (force simp add: part_order_def refl_def)
apply (drule_tac x = "f (x) " and x1 = "f (sa) " in bspec [THEN bspec])
apply (drule_tac [3] x = "g1 (x) " and x1 = "g (sa) " and P2 = "λx y. P (x,y)  d (x,y)  t" for P d in bspec [THEN bspec])
apply auto
apply (rule_tac b = "h (f (sa), g (sa))" and A = C in trans_onD)
apply (auto simp add: part_order_def)
done

(**  This general result is used to prove Follows Un, munion, etc. **)
lemma imp_LeadsTo_comp2:
"F  Increasing(A, r, f1)   Increasing(B, s, g);
  j  A. F: {s  state. <j, f(s)>  r} ⟼w {s  state. <j,f1(s)>  r};
  j  B. F: {x  state. <j, g(x)>  s} ⟼w {x  state. <j,g1(x)>  s};
  mono2(A, r, B, s, C, t, h); refl(A,r); refl(B, s); trans[C](t);
  x  state. f(x):A  g1(x):B  f1(x):A g(x):B; k  C
   F:{x  state. <k, h(f(x), g(x))>  t} ⟼w {x  state. <k, h(f1(x), g1(x))>  t}"
apply (rule_tac B = "{x  state. <k, h (f1 (x), g (x))>  t}" in LeadsTo_Trans)
apply (blast intro: imp_LeadsTo_comp_right)
apply (blast intro: imp_LeadsTo_comp_left)
done

(* Follows type *)
lemma Follows_type: "Follows(A, r, f, g)<=program"
  unfolding Follows_def
apply (blast dest: Increasing_type [THEN subsetD])
done

lemma Follows_into_program [TC]: "F  Follows(A, r, f, g)  F  program"
by (blast dest: Follows_type [THEN subsetD])

lemma FollowsD:
"F  Follows(A, r, f, g)
  F  program  (a. a  A)  (x  state. f(x):A  g(x):A)"
  unfolding Follows_def
apply (blast dest: IncreasingD)
done

lemma Follows_constantI:
 "F  program; c  A; refl(A, r)  F  Follows(A, r, λx. c, λx. c)"
apply (unfold Follows_def, auto)
apply (auto simp add: refl_def)
done

lemma subset_Follows_comp:
"mono1(A, r, B, s, h); refl(A, r); trans[B](s)
    Follows(A, r, f, g)  Follows(B, s,  h comp f, h comp g)"
apply (unfold Follows_def, clarify)
apply (frule_tac f = g in IncreasingD)
apply (frule_tac f = f in IncreasingD)
apply (rule IntI)
apply (rule_tac [2] h = h in imp_LeadsTo_comp)
prefer 5 apply assumption
apply (auto intro: imp_Increasing_comp imp_Always_comp simp del: INT_simps)
done

lemma imp_Follows_comp:
"F  Follows(A, r, f, g);  mono1(A, r, B, s, h); refl(A, r); trans[B](s)
    F  Follows(B, s,  h comp f, h comp g)"
apply (blast intro: subset_Follows_comp [THEN subsetD])
done

(* 2-place monotone operation ∈ this general result is used to prove Follows_Un, Follows_munion *)

(* 2-place monotone operation ∈ this general result is
   used to prove Follows_Un, Follows_munion *)
lemma imp_Follows_comp2:
"F  Follows(A, r, f1, f);  F  Follows(B, s, g1, g);
   mono2(A, r, B, s, C, t, h); refl(A,r); refl(B, s); trans[C](t)
    F  Follows(C, t, λx. h(f1(x), g1(x)), λx. h(f(x), g(x)))"
apply (unfold Follows_def, clarify)
apply (frule_tac f = g in IncreasingD)
apply (frule_tac f = f in IncreasingD)
apply (rule IntI, safe)
apply (rule_tac [3] h = h in imp_Always_comp2)
prefer 5 apply assumption
apply (rule_tac [2] h = h in imp_Increasing_comp2)
prefer 4 apply assumption
apply (rule_tac h = h in imp_Increasing_comp2)
prefer 3 apply assumption
apply simp_all
apply (blast dest!: IncreasingD)
apply (rule_tac h = h in imp_LeadsTo_comp2)
prefer 4 apply assumption
apply auto
  prefer 3 apply (simp add: mono2_def)
apply (blast dest: IncreasingD)+
done


lemma Follows_trans:
     "F  Follows(A, r, f, g);  F  Follows(A,r, g, h);
         trans[A](r)  F  Follows(A, r, f, h)"
apply (frule_tac f = f in FollowsD)
apply (frule_tac f = g in FollowsD)
apply (simp add: Follows_def)
apply (simp add: Always_eq_includes_reachable INT_iff, auto)
apply (rule_tac [2] B = "{s  state. <k, g (s) >  r}" in LeadsTo_Trans)
apply (rule_tac b = "g (x) " in trans_onD)
apply blast+
done

(** Destruction rules for Follows **)

lemma Follows_imp_Increasing_left:
     "F  Follows(A, r, f,g)  F  Increasing(A, r, f)"
by (unfold Follows_def, blast)

lemma Follows_imp_Increasing_right:
     "F  Follows(A, r, f,g)  F  Increasing(A, r, g)"
by (unfold Follows_def, blast)

lemma Follows_imp_Always:
 "F :Follows(A, r, f, g)  F  Always({s  state. <f(s),g(s)>  r})"
by (unfold Follows_def, blast)

lemma Follows_imp_LeadsTo:
 "F  Follows(A, r, f, g); k  A  
  F: {s  state. <k,g(s)>  r } ⟼w {s  state. <k,f(s)>  r}"
by (unfold Follows_def, blast)

lemma Follows_LeadsTo_pfixLe:
     "F  Follows(list(nat), gen_prefix(nat, Le), f, g); k  list(nat)
    F  {s  state. k pfixLe g(s)} ⟼w {s  state. k pfixLe f(s)}"
by (blast intro: Follows_imp_LeadsTo)

lemma Follows_LeadsTo_pfixGe:
     "F  Follows(list(nat), gen_prefix(nat, Ge), f, g); k  list(nat)
    F  {s  state. k pfixGe g(s)} ⟼w {s  state. k pfixGe f(s)}"
by (blast intro: Follows_imp_LeadsTo)

lemma Always_Follows1:
"F  Always({s  state. f(s) = g(s)}); F  Follows(A, r, f, h);
    x  state. g(x):A  F  Follows(A, r, g, h)"
  unfolding Follows_def Increasing_def Stable_def
apply (simp add: INT_iff, auto)
apply (rule_tac [3] C = "{s  state. f(s)=g(s)}"
        and A = "{s  state. <k, h (s)>  r}"
        and A' = "{s  state. <k, f(s)>  r}" in Always_LeadsTo_weaken)
apply (erule_tac A = "{s  state. <k,f(s) >  r}"
           and A' = "{s  state. <k,f(s) >  r}" in Always_Constrains_weaken)
apply auto
apply (drule Always_Int_I, assumption)
apply (erule_tac A = "{s  state. f(s)=g(s)}  {s  state. <f(s), h(s)>  r}"
       in Always_weaken)
apply auto
done


lemma Always_Follows2:
"F  Always({s  state. g(s) = h(s)});
  F  Follows(A, r, f, g); x  state. h(x):A  F  Follows(A, r, f, h)"
  unfolding Follows_def Increasing_def Stable_def
apply (simp add: INT_iff, auto)
apply (rule_tac [3] C = "{s  state. g (s) =h (s) }"
            and A = "{s  state. <k, g (s) >  r}"
            and A' = "{s  state. <k, f (s) >  r}" in Always_LeadsTo_weaken)
apply (erule_tac A = "{s  state. <k, g(s)>  r}"
         and A' = "{s  state. <k, g(s)>  r}" in Always_Constrains_weaken)
apply auto
apply (drule Always_Int_I, assumption)
apply (erule_tac A = "{s  state. g(s)=h(s)}  {s  state. <f(s), g(s)>  r}"
       in Always_weaken)
apply auto
done

(** Union properties (with the subset ordering) **)

lemma refl_SetLe [simp]: "refl(Pow(A), SetLe(A))"
by (unfold refl_def SetLe_def, auto)

lemma trans_on_SetLe [simp]: "trans[Pow(A)](SetLe(A))"
by (unfold trans_on_def SetLe_def, auto)

lemma antisym_SetLe [simp]: "antisym(SetLe(A))"
by (unfold antisym_def SetLe_def, auto)

lemma part_order_SetLe [simp]: "part_order(Pow(A), SetLe(A))"
by (unfold part_order_def, auto)

lemma increasing_Un:
     "F  Increasing.increasing(Pow(A), SetLe(A), f);
         F  Increasing.increasing(Pow(A), SetLe(A), g)
      F  Increasing.increasing(Pow(A), SetLe(A), λx. f(x)  g(x))"
by (rule_tac h = "(Un)" in imp_increasing_comp2, auto)

lemma Increasing_Un:
     "F  Increasing(Pow(A), SetLe(A), f);
         F  Increasing(Pow(A), SetLe(A), g)
      F  Increasing(Pow(A), SetLe(A), λx. f(x)  g(x))"
by (rule_tac h = "(Un)" in imp_Increasing_comp2, auto)

lemma Always_Un:
     "F  Always({s  state. f1(s)  f(s)});
     F  Always({s  state. g1(s)  g(s)})
       F  Always({s  state. f1(s)  g1(s)  f(s)  g(s)})"
by (simp add: Always_eq_includes_reachable, blast)

lemma Follows_Un:
"F  Follows(Pow(A), SetLe(A), f1, f);
     F  Follows(Pow(A), SetLe(A), g1, g)
      F  Follows(Pow(A), SetLe(A), λs. f1(s)  g1(s), λs. f(s)  g(s))"
by (rule_tac h = "(Un)" in imp_Follows_comp2, auto)

(** Multiset union properties (with the MultLe ordering) **)

lemma refl_MultLe [simp]: "refl(Mult(A), MultLe(A,r))"
by (unfold MultLe_def refl_def, auto)

lemma MultLe_refl1 [simp]:
 "multiset(M); mset_of(M)<=A  M, M  MultLe(A, r)"
  unfolding MultLe_def id_def lam_def
apply (auto simp add: Mult_iff_multiset)
done

lemma MultLe_refl2 [simp]: "M  Mult(A)  M, M  MultLe(A, r)"
by (unfold MultLe_def id_def lam_def, auto)


lemma trans_on_MultLe [simp]: "trans[Mult(A)](MultLe(A,r))"
  unfolding MultLe_def trans_on_def
apply (auto intro: trancl_trans simp add: multirel_def)
done

lemma MultLe_type: "MultLe(A, r)<= (Mult(A) * Mult(A))"
apply (unfold MultLe_def, auto)
apply (drule multirel_type [THEN subsetD], auto)
done

lemma MultLe_trans:
     "M,K  MultLe(A,r); K,N  MultLe(A,r)  M,N  MultLe(A,r)"
apply (cut_tac A=A in trans_on_MultLe)
apply (drule trans_onD, assumption)
apply (auto dest: MultLe_type [THEN subsetD])
done

lemma part_order_imp_part_ord:
     "part_order(A, r)  part_ord(A, r-id(A))"
  unfolding part_order_def part_ord_def
apply (simp add: refl_def id_def lam_def irrefl_def, auto)
apply (simp (no_asm) add: trans_on_def)
apply auto
apply (blast dest: trans_onD)
apply (simp (no_asm_use) add: antisym_def)
apply auto
done

lemma antisym_MultLe [simp]:
  "part_order(A, r)  antisym(MultLe(A,r))"
  unfolding MultLe_def antisym_def
apply (drule part_order_imp_part_ord, auto)
apply (drule irrefl_on_multirel)
apply (frule multirel_type [THEN subsetD])
apply (drule multirel_trans)
apply (auto simp add: irrefl_def)
done

lemma part_order_MultLe [simp]:
     "part_order(A, r)   part_order(Mult(A), MultLe(A, r))"
apply (frule antisym_MultLe)
apply (auto simp add: part_order_def)
done

lemma empty_le_MultLe [simp]:
"multiset(M); mset_of(M)<= A  0, M  MultLe(A, r)"
  unfolding MultLe_def
apply (case_tac "M=0")
apply (auto simp add: FiniteFun.intros)
apply (subgoal_tac "<0 +# 0, 0 +# M>  multirel (A, r - id (A))")
apply (rule_tac [2] one_step_implies_multirel)
apply (auto simp add: Mult_iff_multiset)
done

lemma empty_le_MultLe2 [simp]: "M  Mult(A)  0, M  MultLe(A, r)"
by (simp add: Mult_iff_multiset)

lemma munion_mono:
"M, N  MultLe(A, r); K, L  MultLe(A, r) 
  <M +# K, N +# L>  MultLe(A, r)"
  unfolding MultLe_def
apply (auto intro: munion_multirel_mono1 munion_multirel_mono2
       munion_multirel_mono multiset_into_Mult simp add: Mult_iff_multiset)
done

lemma increasing_munion:
     "F  Increasing.increasing(Mult(A), MultLe(A,r), f);
         F  Increasing.increasing(Mult(A), MultLe(A,r), g)
      F  Increasing.increasing(Mult(A),MultLe(A,r), λx. f(x) +# g(x))"
by (rule_tac h = munion in imp_increasing_comp2, auto)

lemma Increasing_munion:
     "F  Increasing(Mult(A), MultLe(A,r), f);
         F  Increasing(Mult(A), MultLe(A,r), g)
      F  Increasing(Mult(A),MultLe(A,r), λx. f(x) +# g(x))"
by (rule_tac h = munion in imp_Increasing_comp2, auto)

lemma Always_munion:
"F  Always({s  state. <f1(s),f(s)>  MultLe(A,r)});
          F  Always({s  state. <g1(s), g(s)>  MultLe(A,r)});
  x  state. f1(x):Mult(A)f(x):Mult(A)  g1(x):Mult(A)  g(x):Mult(A)
       F  Always({s  state. <f1(s) +# g1(s), f(s) +# g(s)>  MultLe(A,r)})"
apply (rule_tac h = munion in imp_Always_comp2, simp_all)
apply (blast intro: munion_mono, simp_all)
done

lemma Follows_munion:
"F  Follows(Mult(A), MultLe(A, r), f1, f);
    F  Follows(Mult(A), MultLe(A, r), g1, g)
   F  Follows(Mult(A), MultLe(A, r), λs. f1(s) +# g1(s), λs. f(s) +# g(s))"
by (rule_tac h = munion in imp_Follows_comp2, auto)

(** Used in ClientImp **)

lemma Follows_msetsum_UN:
"f. i  I. F  Follows(Mult(A), MultLe(A, r), f'(i), f(i));
  s. i  I. multiset(f'(i, s))  mset_of(f'(i, s))<=A 
                        multiset(f(i, s))  mset_of(f(i, s))<=A ;
   Finite(I); F  program
         F  Follows(Mult(A),
                        MultLe(A, r), λx. msetsum(λi. f'(i, x), I, A),
                                      λx. msetsum(λi. f(i,  x), I, A))"
apply (erule rev_mp)
apply (drule Finite_into_Fin)
apply (erule Fin_induct)
apply (simp (no_asm_simp))
apply (rule Follows_constantI)
apply (simp_all (no_asm_simp) add: FiniteFun.intros)
apply auto
apply (rule Follows_munion, auto)
done

end