refactor(library/theories/topology/approaches): rename 'filterlim' to 'tendsto' etc., and general cleaning

This commit is contained in:
Jeremy Avigad 2016-05-13 18:23:09 -04:00 committed by Leonardo de Moura
parent 057935d844
commit b8c230a55d
4 changed files with 523 additions and 504 deletions

View file

@ -101,6 +101,11 @@ xs !mem_singleton
proposition singleton_subset_iff {X : Type} (x : X) (s : set X) : '{x} ⊆ s ↔ x ∈ s :=
iff.intro mem_of_singleton_subset singleton_subset_of_mem
theorem singleton_subset {X : Type} {a : X} {s : set X} (H : a ∈ s) : '{a} ⊆ s :=
take b, suppose b ∈ '{a},
have b = a, from eq_of_mem_singleton this,
show b ∈ s, by rewrite this; assumption
lemma inter_eq_inter_left {X : Type} {s t u : set X} (H₁ : s ∩ t ⊆ u) (H₂ : s ∩ u ⊆ t) :
s ∩ t = s ∩ u :=
eq_of_subset_of_subset
@ -329,6 +334,125 @@ theorem inv_fun_spec' {X Y : Type} {f : X → Y} {a : set X} {dflt : X} {x : X}
inv_fun f a dflt (f x) ∈ a :=
and.left (inv_fun_pos (exists.intro x (and.intro xa rfl)))
-- TODO: move to data.set.filter
namespace filter
protected theorem le_iff {X : Type} (F₁ F₂ : filter X) : F₁ ≤ F₂ ↔ F₂ ⊆ F₁ := iff.refl _
-- TODO: change names of fields in filter
-- TODO: reorder hypotheses in eventually_of_le, and change "le" to "ge"
-- TODO: fix eventually_inf: implicit argument, and use implication
-- TODO: set: add spaces around ∀₀ x ∈ s and ∃₀ x ∈ s
theorem eventually_inf_left {X : Type} {P : X → Prop} {F₁ : filter X} (F₂ : filter X)
(H : eventually P F₁) : eventually P (inf F₁ F₂) :=
eventually_of_le H !inf_le_left
theorem eventually_inf_right {X : Type} {P : X → Prop} (F₁ : filter X) {F₂ : filter X}
(H : eventually P F₂) : eventually P (inf F₁ F₂) :=
eventually_of_le H !inf_le_right
theorem eventually_Inf {X : Type} {P : X → Prop} {S : set (filter X)} {F : filter X} (FS : F ∈ S)
(H : eventually P F) : eventually P (Inf S) :=
eventually_of_le H (Inf_le FS)
-- TODO: replace definition of Inf with this
definition Inf' {X : Type} (S : set (filter X)) : filter X :=
⦃ filter,
sets := { s | ∃ T : set (set X), finite T ∧ T ⊆ ( F ∈ S, sets F) ∧ ⋂₀ T ⊆ s},
univ_mem_sets := abstract
have H : (⋂₀ ∅) ⊆ @univ X, by rewrite sInter_empty; apply subset.refl,
exists.intro ∅ (and.intro !finite_empty (and.intro (empty_subset _) H))
end,
inter_closed := abstract
take s t, assume Hs Ht,
obtain Ts finTs Tssub ITs, from Hs,
obtain Tt finTt Ttsub ITt, from Ht,
have H1 : finite (Ts Tt), proof finite_union Ts Tt qed,
have H2 : Ts Tt ⊆ ( F ∈ S, sets F), from union_subset Tssub Ttsub,
have H3 : ⋂₀ (Ts Tt) ⊆ s ∩ t,
begin
rewrite sInter_union, apply subset_inter,
{exact subset.trans (inter_subset_left _ _) ITs},
exact subset.trans (inter_subset_right _ _) ITt
end,
exists.intro _ (and.intro H1 (and.intro H2 H3))
end,
is_mono := abstract
take s t ssubt Hs,
obtain T finT Tsub IT, from Hs,
exists.intro T (and.intro finT (and.intro Tsub (subset.trans IT ssubt)))
end
theorem sets_Inf' {A : Type} (S : set (filter A)) :
sets (Inf' S) = { s | ∃ T : set (set A), finite T ∧ T ⊆ ( F ∈ S, sets F) ∧ ⋂₀ T ⊆ s} :=
rfl
theorem sInter_mem_of_finite {A : Type} {F : filter A} {T : set (set A)} (finT : finite T)
(Tsub : T ⊆ sets F) : ⋂₀ T ∈ sets F :=
begin
induction finT with a T finT aninT ih,
{rewrite sInter_empty, apply filter.univ_mem_sets},
rewrite sInter_insert, apply filter.inter_closed,
show a ∈ sets F, from Tsub (mem_insert a T),
show ⋂₀ T ∈ sets F, from ih (subset.trans (subset_insert _ _) Tsub)
end
theorem le_Inf' {A : Type} {F : filter A} {S : set (filter A)} (H : ∀₀ G ∈ S, F ≤ G) :
F ≤ Inf' S :=
filter.le_of_subset
(take s, suppose s ∈ sets (Inf' S),
obtain (T : set (set A)) finT (Tsub : T ⊆ ( G ∈ S, sets G)) (IT : ⋂₀ T ⊆ s), from this,
have T ⊆ sets F, from subset.trans Tsub (bUnion_subset H),
have ⋂₀ T ∈ sets F, from sInter_mem_of_finite finT this,
show s ∈ sets F, from filter.is_mono _ IT this)
theorem Inf'_le {A : Type} {F : filter A} {S : set (filter A)} (FS : F ∈ S) :
Inf' S ≤ F :=
filter.le_of_subset
(take s, suppose s ∈ sets F,
have '{s} ⊆ G ∈ S, sets G, from singleton_subset (mem_bUnion FS this),
exists.intro '{s} (and.intro _
(and.intro this (by rewrite sInter_singleton; apply subset.refl))))
theorem Inf_eq_Inf' {A : Type} (S : set (filter A)) : Inf S = Inf' S :=
le.antisymm (le_Inf' (λ F FS, Inf_le FS)) (le_Inf (λ F FS, Inf'_le FS))
theorem exists_eventually_of_eventually_Inf {A : Type} {P : A → Prop} {F : filter A}
{S : set (filter A)} (FS : F ∈ S) (H' : eventually P (Inf S))
(H : ∀₀ F₁ ∈ S, ∀₀ F₂ ∈ S, ∃₀ F ∈ S, F ≤ inf F₁ F₂) :
∃₀ F ∈ S, eventually P F :=
have P ∈ Inf' S, by rewrite -Inf_eq_Inf'; apply H',
have ∃ T : set (set A), finite T ∧ T ⊆ ( F ∈ S, sets F) ∧ ⋂₀ T ⊆ P,
by rewrite sets_Inf' at this; exact this,
obtain T finT Tsub ITP, from this,
have ∃₀ F ∈ S, ⋂₀ T ∈ F,
begin
clear ITP,
induction finT with s T finT sninT ih,
{exact exists.intro F (and.intro FS (by rewrite sInter_empty; apply filter.univ_mem_sets))},
have ∃₀ F ∈ S, ⋂₀ T ∈ F, from ih (subset.trans (subset_insert _ _) Tsub),
cases this with F₁ H₁,
cases H₁ with F₁S ITF₁,
have s ∈ ( F ∈ S, sets F), from Tsub !mem_insert,
cases this with F₂ H₂,
cases H₂ with F₂S sF₂,
cases H F₁S F₂S with F' HF',
cases HF' with F'S F'le,
existsi F', split, exact F'S,
show ⋂₀ insert s T ∈ sets F',
begin
rewrite sInter_insert, apply filter.inter_closed,
show s ∈ sets F', from filter.subset_of_le (le.trans F'le !inf_le_right) sF₂,
show ⋂₀ T ∈ sets F', from filter.subset_of_le (le.trans F'le !inf_le_left) ITF₁,
end
end,
obtain F FS IT, from this,
exists.intro F (and.intro FS (filter.is_mono _ ITP IT))
end filter
end set

View file

@ -1,503 +0,0 @@
/-
Copyright (c) 2016 Jacob Gross. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Jacob Gross, Jeremy Avigad
Limits, following Hölzl, Immler, and Huffman, "Type classes and filters for mathematical
analysis in Isabelle/HOL".
We are experimenting with notation like this:
f ⟶ y [at x]
... which is actually a composition of two notations: the first describes the function and
the target filter, and the second desribes the source filter.
Variations on the first can include:
f ⟶ ∞
f ⟶ -∞
f ⟶+ y convergence from the right
f ⟶- y convergence from the left
f ⟶ y within s convergences within a set
Variations on the second can include:
[at ∞]
[at -∞]
[at x]
[at+ x]
[at- x]
[at x within s]
When written consecutively, we get filterlim f F₂ F₁, where the first notation gives f and F₂,
and the second gives F₁.
Note that lambdas are needed for functions given by expressions:
(λ x, x^2) ⟶ 4 [at 2]
(λ x, x^2) ⟶ ∞ [at -∞]
If making `[at` a token is problematic, we can add a tick mark or something.
-/
import data.set data.nat algebra.order algebra.interval
open set function set.filter interval
-- TODO: move to data/set/basic
theorem singleton_subset {X : Type} {a : X} {s : set X} (H : a ∈ s) : '{a} ⊆ s :=
take b, suppose b ∈ '{a},
have b = a, from eq_of_mem_singleton this,
show b ∈ s, by rewrite this; assumption
theorem mem_of_singleton_subset {X : Type} {a : X} {s : set X} (H : '{a} ⊆ s) : a ∈ s :=
H (mem_singleton a)
theorem singleton_subset_iff {X : Type} (a : X) (s : set X) : '{a} ⊆ s ↔ a ∈ s :=
iff.intro mem_of_singleton_subset singleton_subset
-- TODO: move these to data/set/filter
namespace filter
protected theorem le_iff {X : Type} (F₁ F₂ : filter X) : F₁ ≤ F₂ ↔ F₂ ⊆ F₁ := iff.refl _
end filter
-- TODO: change names of fields in filter
-- TODO: reorder hypotheses in eventually_of_le, and change "le" to "ge"
-- TODO: fix eventually_inf: implicit argument, and use implication
-- TODO: set: add spaces around ∀₀ x ∈ s and ∃₀ x ∈ s
theorem eventually_inf_left {X : Type} {P : X → Prop} {F₁ : filter X} (F₂ : filter X)
(H : eventually P F₁) : eventually P (inf F₁ F₂) :=
eventually_of_le H !inf_le_left
theorem eventually_inf_right {X : Type} {P : X → Prop} (F₁ : filter X) {F₂ : filter X}
(H : eventually P F₂) : eventually P (inf F₁ F₂) :=
eventually_of_le H !inf_le_right
theorem eventually_Inf {X : Type} {P : X → Prop} {S : set (filter X)} {F : filter X} (FS : F ∈ S)
(H : eventually P F) : eventually P (Inf S) :=
eventually_of_le H (Inf_le FS)
-- TODO: replace definition of Inf with this
definition Inf' {X : Type} (S : set (filter X)) : filter X :=
⦃ filter,
sets := { s | ∃ T : set (set X), finite T ∧ T ⊆ ( F ∈ S, sets F) ∧ ⋂₀ T ⊆ s},
univ_mem_sets := abstract
have H : (⋂₀ ∅) ⊆ @univ X, by rewrite sInter_empty; apply subset.refl,
exists.intro ∅ (and.intro !finite_empty (and.intro (empty_subset _) H))
end,
inter_closed := abstract
take s t, assume Hs Ht,
obtain Ts finTs Tssub ITs, from Hs,
obtain Tt finTt Ttsub ITt, from Ht,
have H1 : finite (Ts Tt), proof finite_union Ts Tt qed,
have H2 : Ts Tt ⊆ ( F ∈ S, sets F), from union_subset Tssub Ttsub,
have H3 : ⋂₀ (Ts Tt) ⊆ s ∩ t,
begin
rewrite sInter_union, apply subset_inter,
{exact subset.trans (inter_subset_left _ _) ITs},
exact subset.trans (inter_subset_right _ _) ITt
end,
exists.intro _ (and.intro H1 (and.intro H2 H3))
end,
is_mono := abstract
take s t ssubt Hs,
obtain T finT Tsub IT, from Hs,
exists.intro T (and.intro finT (and.intro Tsub (subset.trans IT ssubt)))
end
theorem sets_Inf' {A : Type} (S : set (filter A)) :
sets (Inf' S) = { s | ∃ T : set (set A), finite T ∧ T ⊆ ( F ∈ S, sets F) ∧ ⋂₀ T ⊆ s} :=
rfl
theorem sInter_mem_of_finite {A : Type} {F : filter A} {T : set (set A)} (finT : finite T)
(Tsub : T ⊆ sets F) : ⋂₀ T ∈ sets F :=
begin
induction finT with a T finT aninT ih,
{rewrite sInter_empty, apply filter.univ_mem_sets},
rewrite sInter_insert, apply filter.inter_closed,
show a ∈ sets F, from Tsub (mem_insert a T),
show ⋂₀ T ∈ sets F, from ih (subset.trans (subset_insert _ _) Tsub)
end
theorem le_Inf' {A : Type} {F : filter A} {S : set (filter A)} (H : ∀₀ G ∈ S, F ≤ G) :
F ≤ Inf' S :=
filter.le_of_subset
(take s, suppose s ∈ sets (Inf' S),
obtain (T : set (set A)) finT (Tsub : T ⊆ ( G ∈ S, sets G)) (IT : ⋂₀ T ⊆ s), from this,
have T ⊆ sets F, from subset.trans Tsub (bUnion_subset H),
have ⋂₀ T ∈ sets F, from sInter_mem_of_finite finT this,
show s ∈ sets F, from filter.is_mono _ IT this)
theorem Inf'_le {A : Type} {F : filter A} {S : set (filter A)} (FS : F ∈ S) :
Inf' S ≤ F :=
filter.le_of_subset
(take s, suppose s ∈ sets F,
have '{s} ⊆ G ∈ S, sets G, from singleton_subset (mem_bUnion FS this),
exists.intro '{s} (and.intro _
(and.intro this (by rewrite sInter_singleton; apply subset.refl))))
theorem Inf_eq_Inf' {A : Type} (S : set (filter A)) : Inf S = Inf' S :=
le.antisymm (le_Inf' (λ F FS, Inf_le FS)) (le_Inf (λ F FS, Inf'_le FS))
theorem exists_eventually_of_eventually_Inf {A : Type} {P : A → Prop} {F : filter A}
{S : set (filter A)} (FS : F ∈ S) (H : ∀₀ F₁ ∈ S, ∀₀ F₂ ∈ S, ∃₀ F ∈ S, F ≤ inf F₁ F₂)
(H' : eventually P (Inf S)) :
∃₀ F ∈ S, eventually P F :=
have P ∈ Inf' S, by rewrite -Inf_eq_Inf'; apply H',
have ∃ T : set (set A), finite T ∧ T ⊆ ( F ∈ S, sets F) ∧ ⋂₀ T ⊆ P,
by rewrite sets_Inf' at this; exact this,
obtain T finT Tsub ITP, from this,
have ∃₀ F ∈ S, ⋂₀ T ∈ F,
begin
clear ITP,
induction finT with s T finT sninT ih,
{exact exists.intro F (and.intro FS (by rewrite sInter_empty; apply filter.univ_mem_sets))},
have ∃₀ F ∈ S, ⋂₀ T ∈ F, from ih (subset.trans (subset_insert _ _) Tsub),
cases this with F₁ H₁,
cases H₁ with F₁S ITF₁,
have s ∈ ( F ∈ S, sets F), from Tsub !mem_insert,
cases this with F₂ H₂,
cases H₂ with F₂S sF₂,
cases H F₁S F₂S with F' HF',
cases HF' with F'S F'le,
existsi F', split, exact F'S,
show ⋂₀ insert s T ∈ sets F',
begin
rewrite sInter_insert, apply filter.inter_closed,
show s ∈ sets F', from filter.subset_of_le (le.trans F'le !inf_le_right) sF₂,
show ⋂₀ T ∈ sets F', from filter.subset_of_le (le.trans F'le !inf_le_left) ITF₁,
end
end,
obtain F FS IT, from this,
exists.intro F (and.intro FS (filter.is_mono _ ITP IT))
-- the file proper starts here
variables {X Y Z : Type}
/- mapfilter -/
section
local attribute mem [reducible]
definition mapfilter (f : X → Y) (F : filter X) : filter Y :=
⦃ filter,
sets := λ s, (f '- s) ∈ F,
univ_mem_sets := abstract !univ_mem_sets end,
inter_closed := abstract take a b, assume Ha Hb, !inter_closed Ha Hb end,
is_mono := abstract take a b, assume Hab Ha, !is_mono (λ x H, Hab (f x) H) Ha end
end
-- characterize mapfilter in terms of set membership
theorem mem_mapfilter_iff (s : set Y) (F : filter X) (f : X → Y) :
s ∈ mapfilter f F ↔ (f '- s) ∈ F :=
!iff.refl
theorem mem_mapfilter {s : set Y} {F : filter X} {f : X → Y} (H : f '- s ∈ F) :
s ∈ mapfilter f F := H
theorem preimage_mem_filter_of_mem_mapfilter {s : set Y} {F : filter X} {f : X → Y}
(H : s ∈ mapfilter f F) :
f '- s ∈ F := H
-- characterize mapfilter in terms of eventually and predicates
theorem eventually_mapfilter_iff (P : Y → Prop) (F : filter X) (f : X → Y) :
eventually P (mapfilter f F) ↔ eventually (λ x, P (f x)) F :=
!iff.refl
theorem eventually_mapfilter {P : Y → Prop} {F : filter X} {f : X → Y}
(H : eventually (λ x, P (f x)) F) :
eventually P (mapfilter f F) := H
theorem eventually_of_eventually_mapfilter {P : Y → Prop} {F : filter X} {f : X → Y}
(H : eventually P (mapfilter f F)) :
eventually (λ x, P (f x)) F:= H
-- other properties
theorem mapfilter_id (F : filter Y) : mapfilter (λ x, x) F = F := filter.eq rfl
theorem mapfilter_comp (F : filter X) (g : Y → Z) (f : X → Y) :
mapfilter (g ∘ f) F = mapfilter g (mapfilter f F) := rfl
theorem mapfilter_le_mapfilter {F F' : filter X} (f : X → Y) (H : F ≤ F') :
mapfilter f F ≤ mapfilter f F' :=
begin
rewrite filter.le_iff, intro s, rewrite *mem_mapfilter_iff,
apply filter.subset_of_le H
end
-- alternative proof: take s, assume H', by unfold mem; exact H H'
theorem mapfilter_bot (f : X → Y) : mapfilter f ⊥ = ⊥ :=
filter.eq (by rewrite *filter.bot_eq)
theorem mapfilter_sup (f : X → Y) (F F' : filter X) :
mapfilter f (sup F F') = sup (mapfilter f F) (mapfilter f F') :=
filter.eq (ext (take x, !iff.refl))
theorem mapfilter_inf (f : X → Y) (F F' : filter X) :
mapfilter f (inf F F') ≤ inf (mapfilter f F) (mapfilter f F') :=
le_inf (mapfilter_le_mapfilter f !inf_le_left) (mapfilter_le_mapfilter f !inf_le_right)
/- limits -/
definition filterlim (f : X → Y) (F₂ : filter Y) (F₁ : filter X) : Prop := mapfilter f F₁ ≤ F₂
theorem filterlim_of_forall_eventually {F₂ : filter Y} {F₁ : filter X} {f : X → Y}
(H : ∀ P, eventually P F₂ → eventually P (mapfilter f F₁)) :
filterlim f F₂ F₁ := H
theorem eventually_mapfilter_of_filterlim {F₂ : filter Y} {F₁ : filter X} {f : X → Y}
{P : Y → Prop} (H : filterlim f F₂ F₁) (H' : eventually P F₂) :
eventually P (mapfilter f F₁) := H H'
theorem filterlim_iff (F₂ : filter Y) (F₁ : filter X) (f : X → Y) :
filterlim f F₂ F₁ ↔ (∀ P, eventually P F₂ → eventually P (mapfilter f F₁)) :=
iff.refl _
theorem filterlim_iff' (F₂ : filter Y) (F₁ : filter X) (f : X → Y) :
filterlim f F₂ F₁ ↔ (∀ P, eventually P F₂ → eventually (λ x, P (f x)) F₁) :=
iff.refl _
theorem filterlim_comp {f : X → Y} {g : Y → Z} {F₂ : filter Y} {F₁ : filter X} {F₃ : filter Z}
(Hf : filterlim f F₂ F₁) (Hg : filterlim g F₃ F₂) : filterlim (g ∘ f) F₃ F₁ :=
filterlim_of_forall_eventually
(take P, suppose eventually P F₃,
have eventually P (mapfilter g F₂), from Hg this,
have eventually (λ y, P (g y)) F₂, from eventually_of_eventually_mapfilter this,
have eventually (λ y, P (g y)) (mapfilter f F₁), from Hf this,
show eventually P (mapfilter g (mapfilter f F₁)), from eventually_mapfilter this)
theorem filterlim_of_ge {f : X → Y} {F₂ F₂' : filter Y} {F₁ : filter X}
(HF₂ : F₂' ≥ F₂) (H : filterlim f F₂ F₁) : filterlim f F₂' F₁ :=
take P, suppose eventually P F₂',
have eventually P F₂, from eventually_of_le this HF₂,
show eventually P (mapfilter f F₁), from H this
theorem filterlim_of_le {f : X → Y} {F₂ : filter Y} {F₁ F₁' : filter X}
(HF₁ : F₁' ≤ F₁) (H : filterlim f F₂ F₁) : filterlim f F₂ F₁' :=
take P, suppose eventually P F₂,
have eventually P (mapfilter f F₁), from H this,
show eventually P (mapfilter f F₁'), from eventually_of_le this (mapfilter_le_mapfilter _ HF₁)
theorem filterlim_of_ge_of_le {f : X → Y} {F₂ F₂' : filter Y} {F₁ F₁' : filter X}
(HF₂ : F₂' ≥ F₂) (HF₁ : F₁' ≤ F₁) (H : filterlim f F₂ F₁) : filterlim f F₂' F₁' :=
filterlim_of_ge HF₂ (filterlim_of_le HF₁ H)
theorem filterlim_id {F : filter X} : filterlim (λx, x) F F :=
take P, assume H, H
theorem filterlim_inf_left {f : X → Y} {F : filter Y} (F₁ : filter X) {F₂ : filter X}
(H : filterlim f F F₁) :
filterlim f F (inf F₁ F₂) :=
filterlim_of_le !inf_le_left H
theorem filterlim_inf_right {f : X → Y} {F : filter Y} {F₁ : filter X} (F₂ : filter X)
(H : filterlim f F F₂) :
filterlim f F (inf F₁ F₂) :=
filterlim_of_le !inf_le_right H
theorem filterlim_sup_left {f : X → Y} (F₁ : filter Y) {F₂ : filter Y} {F : filter X}
(H : filterlim f F₁ F) :
filterlim f (sup F₁ F₂) F :=
filterlim_of_ge !le_sup_left H
theorem filterlim_sup_right {f : X → Y} {F₁ : filter Y} (F₂ : filter Y) {F : filter X}
(H : filterlim f F₂ F) :
filterlim f (sup F₁ F₂) F :=
filterlim_of_ge !le_sup_right H
theorem filterlim_comp_iff_filterlim_mapfilter (f : X → Y) (g : Y → Z)
(F₁ : filter X) (F₃ : filter Z) :
filterlim (g ∘ f) F₃ F₁ ↔ filterlim g F₃ (mapfilter f F₁) :=
!iff.refl
/- limits on orders -/
section linorderX
variable [linear_strong_order_pair X]
/- at_top filter, for limits as x goes to infinity -/
definition at_top : filter X := Inf ((λ x : X, principal '[x, ∞)) ' univ)
private lemma principal_Ici_le_principal_Ici {x₁ x₂ : X} (H : x₁ ≤ x₂) :
(principal '[x₂, ∞)) ≤ (principal '[x₁, ∞)) :=
principal_le_principal (show '[x₂, ∞) ⊆ '[x₁, ∞), from λ y Hy, le.trans H Hy)
theorem eventually_at_top {P : X → Prop} {x : X} (H : ∀ y, y ≥ x → P y) : eventually P at_top :=
have H' : eventually P (principal '[x, ∞)), from eventually_principal H,
have principal '[x, ∞) ∈ (λ x : X, principal '[x, ∞)) ' univ, from mem_image_of_mem _ !mem_univ,
eventually_Inf this H'
theorem exists_forall_ge_imp_of_eventually_at_top {P : X → Prop} [inhabited X]
(H : eventually P at_top) :
∃ x, ∀ y, y ≥ x → P y :=
let S := (λ x : X, principal '[x, ∞)) ' univ in
have H' : ∀₀ F₁ ∈ S, ∀₀ F₂ ∈ S, ∃₀ F ∈ S, F ≤ F₁ ⊓ F₂,
proof
take F₁, suppose F₁ ∈ S, take F₂, suppose F₂ ∈ S,
obtain x₁ [x₁univ F₁eq], from `F₁ ∈ S`,
obtain x₂ [x₂univ F₂eq], from `F₂ ∈ S`,
or.elim (le_or_gt x₁ x₂)
(suppose x₁ ≤ x₂,
have F₂ ≤ F₁, by rewrite [-F₁eq, -F₂eq]; exact principal_Ici_le_principal_Ici this,
exists.intro F₂ (and.intro `F₂ ∈ S` (le_inf this !le.refl)))
(suppose x₂ < x₁,
have x₂ ≤ x₁, from le_of_lt this,
have F₁ ≤ F₂, by rewrite [-F₁eq, -F₂eq]; exact principal_Ici_le_principal_Ici this,
exists.intro F₁ (and.intro `F₁ ∈ S` (le_inf !le.refl this)))
qed,
have principal '[arbitrary X, ∞) ∈ S,
from mem_image_of_mem _ !mem_univ,
have ∃₀ F ∈ S, eventually P F,
from exists_eventually_of_eventually_Inf this H' H,
obtain F [FS ePF], from this,
obtain x [xuniv Feq], from FS,
have P ⊇ '[x, ∞), from subset_of_eventually_principal (eq.subst (eq.symm Feq) ePF),
exists.intro x this
theorem eventually_at_top_iff (P : X → Prop) [inhabited X] :
eventually P at_top ↔ ∃ x, ∀ y, y ≥ x → P y :=
iff.intro exists_forall_ge_imp_of_eventually_at_top
(assume H, obtain x Hx, from H, eventually_at_top Hx)
/- at_bot filter, for limits as x goes to negative infinity -/
definition at_bot : filter X := Inf ((λ x : X, principal '(-∞, x]) ' univ)
private lemma principal_Iic_le_principal_Iic {x₁ x₂ : X} (H : x₁ ≤ x₂) :
(principal '(-∞, x₁]) ≤ (principal '(-∞, x₂]) :=
principal_le_principal (show '(-∞, x₁] ⊆ '(-∞, x₂], from λ y Hy, le.trans Hy H)
theorem eventually_at_bot {P : X → Prop} {x : X} (H : ∀ y, y ≤ x → P y) : eventually P at_bot :=
have H' : eventually P (principal '(-∞, x]), from eventually_principal H,
have principal '(-∞, x] ∈ (λ x : X, principal '(-∞, x]) ' univ, from mem_image_of_mem _ !mem_univ,
eventually_Inf this H'
theorem exists_forall_le_imp_of_eventually_at_bot {P : X → Prop} [inhabited X]
(H : eventually P at_bot) :
∃ x, ∀ y, y ≤ x → P y :=
let S := (λ x : X, principal '(-∞, x]) ' univ in
have H' : ∀₀ F₁ ∈ S, ∀₀ F₂ ∈ S, ∃₀ F ∈ S, F ≤ F₁ ⊓ F₂,
proof
take F₁, suppose F₁ ∈ S, take F₂, suppose F₂ ∈ S,
obtain x₁ [x₁univ F₁eq], from `F₁ ∈ S`,
obtain x₂ [x₂univ F₂eq], from `F₂ ∈ S`,
or.elim (le_or_gt x₁ x₂)
(suppose x₁ ≤ x₂,
have F₁ ≤ F₂, by rewrite [-F₁eq, -F₂eq]; exact principal_Iic_le_principal_Iic this,
exists.intro F₁ (and.intro `F₁ ∈ S` (le_inf !le.refl this)))
(suppose x₂ < x₁,
have x₂ ≤ x₁, from le_of_lt this,
have F₂ ≤ F₁, by rewrite [-F₁eq, -F₂eq]; exact principal_Iic_le_principal_Iic this,
exists.intro F₂ (and.intro `F₂ ∈ S` (le_inf this !le.refl)))
qed,
have principal '(-∞, arbitrary X] ∈ S,
from mem_image_of_mem _ !mem_univ,
have ∃₀ F ∈ S, eventually P F,
from exists_eventually_of_eventually_Inf this H' H,
obtain F [FS ePF], from this,
obtain x [xuniv Feq], from FS,
have P ⊇ '(-∞, x], from subset_of_eventually_principal (eq.subst (eq.symm Feq) ePF),
exists.intro x this
theorem eventually_at_bot_iff (P : X → Prop) [inhabited X] :
eventually P at_bot ↔ ∃ x, ∀ y, y ≤ x → P y :=
iff.intro exists_forall_le_imp_of_eventually_at_bot
(assume H, obtain x Hx, from H, eventually_at_bot Hx)
/- notation -/
notation f ` ⟶ ` ∞ := filterlim f at_top
notation f ` ⟶ ` -∞ := filterlim f at_bot
notation `[at ` `∞]` := at_top
notation `[at ` `-∞]` := at_bot
section examples
variable [linear_strong_order_pair Y]
variable f : X → Y
-- check f ⟶ ∞ [at ∞]
-- check f ⟶ ∞ [at -∞]
-- check f ⟶ -∞ [at -∞]
-- check (λ x : X, x) ⟶ ∞ [at ∞]
end examples
end linorderX
/- limits at top and at bot -/
section linorderY
variable [linear_strong_order_pair Y]
theorem filterlim_at_top [inhabited Y] {f : X → Y} {F : filter X}
(H : ∀ y, eventually (λ x, f x ≥ y) F) :
(f ⟶ ∞) F :=
filterlim_of_forall_eventually
(take P,
suppose eventually P at_top,
obtain z (Hz : ∀ y, y ≥ z → P y), from exists_forall_ge_imp_of_eventually_at_top this,
have ∀ x, f x ≥ z → P (f x), from take x, Hz (f x),
have eventually (λ x, P (f x)) F, from eventually_mono (H z) this,
show eventually P (mapfilter f F), from eventually_mapfilter this)
theorem eventually_ge_of_filterlim_at_top {f : X → Y} {F : filter X} (H : (f ⟶ ∞) F) (y : Y) :
eventually (λ x, f x ≥ y) F :=
have eventually (λ x, x ≥ y) at_top, from eventually_at_top (take x, suppose x ≥ y, this),
have eventually (λ x, x ≥ y) (mapfilter f F), from H this,
show eventually (λ x, f x ≥ y) F, from eventually_of_eventually_mapfilter this
theorem filterlim_at_top_iff [inhabited Y] (f : X → Y) (F : filter X) :
(f ⟶ ∞) F ↔ ∀ y, eventually (λ x, f x ≥ y) F :=
iff.intro eventually_ge_of_filterlim_at_top filterlim_at_top
theorem filterlim_at_top_of_eventually_ge [inhabited Y] {f g : X → Y} {F : filter X}
(H : eventually (λ x, f x ≥ g x) F) (H' : (g ⟶ ∞) F) :
(f ⟶ ∞) F :=
filterlim_at_top (take y,
have eventually (λ x, g x ≥ y) F, from eventually_ge_of_filterlim_at_top H' y,
show eventually (λ x, f x ≥ y) F, from
eventually_mpr H (eventually_mpr this (eventually_of_forall _ (take x, le.trans))))
theorem filterlim_at_bot [inhabited Y] {f : X → Y} {F : filter X}
(H : ∀ y, eventually (λ x, f x ≤ y) F) :
(f ⟶ -∞) F :=
filterlim_of_forall_eventually
(take P,
suppose eventually P at_bot,
obtain z (Hz : ∀ y, y ≤ z → P y), from exists_forall_le_imp_of_eventually_at_bot this,
have ∀ x, f x ≤ z → P (f x), from take x, Hz (f x),
have eventually (λ x, P (f x)) F, from eventually_mono (H z) this,
show eventually P (mapfilter f F), from eventually_mapfilter this)
theorem eventually_ge_of_filterlim_at_bot {f : X → Y} {F : filter X} (H : (f ⟶ -∞) F) (y : Y) :
eventually (λ x, f x ≤ y) F :=
have eventually (λ x, x ≤ y) at_bot, from eventually_at_bot (take x, suppose x ≤ y, this),
have eventually (λ x, x ≤ y) (mapfilter f F), from H this,
show eventually (λ x, f x ≤ y) F, from eventually_of_eventually_mapfilter this
theorem filterlim_at_bot_iff [inhabited Y] (f : X → Y) (F : filter X) :
(f ⟶ -∞) F ↔ ∀ y, eventually (λ x, f x ≤ y) F :=
iff.intro eventually_ge_of_filterlim_at_bot filterlim_at_bot
theorem filterlim_at_bot_of_eventually_ge [inhabited Y] {f g : X → Y} {F : filter X}
(H : eventually (λ x, f x ≤ g x) F) (H' : (g ⟶ -∞) F) :
(f ⟶ -∞) F :=
filterlim_at_bot (take y,
have eventually (λ x, g x ≤ y) F, from eventually_ge_of_filterlim_at_bot H' y,
show eventually (λ x, f x ≤ y) F, from
eventually_mpr H (eventually_mpr this (eventually_of_forall _
(take x H₁ H₂, le.trans H₂ H₁))))
end linorderY

View file

@ -0,0 +1,398 @@
/-
Copyright (c) 2016 Jacob Gross. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Jacob Gross, Jeremy Avigad
Limits, following Hölzl, Immler, and Huffman, "Type classes and filters for mathematical analysis
in Isabelle/HOL".
The central notion in this file is
tendsto f F₂ F₁
which says that f "tends to" filter F₂, as the input "tends to" filter F₁. If f : X → Y, F₁ is a
filter on X, and F₂ is a filter on Y.
We can then use notation like this:
f ⟶ y [at x]
... which is actually a composition of two notations: the first describes the function and the
target filter, and the second desribes the source filter.
We define the following filters:
nhds x := neighborhoods of x
[at x within s] := at_within x s := neighborhoods of x intersected with s \ '{x}
[at x] := at_elt x := [at x within univ]
[at x-] := at_left x := [at x within '(-∞, x)
[at x+] := at_right x := [at x within '(x, ∞)
[at ∞] := at_infty := "neighborhoods" of ∞
[at -∞] := at_ninfty := "neighborhoods" of -∞
We also define the following abbreviations and notatations for modes of convergence:
f ⟶ y := approaches f y := tendsto f (nhds y)
f ⟶ ∞ := approaches_infty f := tendsto f at_infty
f ⟶ -∞ := approaches_ninfty := tendsto f at_ninfty
Thus, for example, "f ⟶ y [at x]" denotes "tendsto f (nhds y) (at_elt x)".
Note that lambdas are needed for functions given by expressions, for example:
(λ x, x^2) ⟶ 4 [at 2]
(λ x, x^2) ⟶ ∞ [at -∞]
Using the theorems can take some getting used to. "tendsto f F₂ F₁" is equivalent to
∀ P : Y → Prop, eventually P F₂ → eventually (λ x, P (f x)) F₁
In terms of sets, this is equivalent to
∀ s : set Y, s ∈ sets F₂ → f '- s ∈ sets F₁
so one option is to use "tendsto_intro" "tendsto_elim" and "tendsto_iff" to unpack meanings
in terms of "eventually", and then use the intro and elim rules of "eventually". For specific
topologies -- for example, those given by a metric space or norm -- see the specialized
intro and elim rules for eventually in those files.
For "approaches", "approaches_infty", and "approaches_ninfty", there are specific intro, elim,
and iff rules. Again, see also specific versions for metric spaces, normed spaces, etc.
Note that the filters at_infty and at_ninfty don't rely on topological notions at all, only the
existence of a suitable order.
-/
import data.set data.nat algebra.interval
import theories.move -- TODO: remove after move to Lean 3
open set function set.filter interval
namespace topology
variables {X Y Z : Type}
/- mapfilter -/
section
local attribute mem [reducible]
definition mapfilter (f : X → Y) (F : filter X) : filter Y :=
⦃ filter,
sets := λ s, (f '- s) ∈ F,
univ_mem_sets := abstract !univ_mem_sets end,
inter_closed := abstract take a b, assume Ha Hb, !inter_closed Ha Hb end,
is_mono := abstract take a b, assume Hab Ha, !is_mono (λ x H, Hab (f x) H) Ha end
end
-- characterize mapfilter in terms of set membership
theorem mem_mapfilter_iff (s : set Y) (F : filter X) (f : X → Y) :
s ∈ mapfilter f F ↔ (f '- s) ∈ F :=
!iff.refl
theorem mem_mapfilter {s : set Y} {F : filter X} {f : X → Y} (H : f '- s ∈ F) :
s ∈ mapfilter f F := H
theorem preimage_mem_filter_of_mem_mapfilter {s : set Y} {F : filter X} {f : X → Y}
(H : s ∈ mapfilter f F) :
f '- s ∈ F := H
-- characterize mapfilter in terms of predicates and eventually
theorem eventually_mapfilter_iff (P : Y → Prop) (F : filter X) (f : X → Y) :
eventually P (mapfilter f F) ↔ eventually (λ x, P (f x)) F :=
!iff.refl
theorem eventually_mapfilter {P : Y → Prop} {F : filter X} {f : X → Y}
(H : eventually (λ x, P (f x)) F) :
eventually P (mapfilter f F) := H
theorem eventually_of_eventually_mapfilter {P : Y → Prop} {F : filter X} {f : X → Y}
(H : eventually P (mapfilter f F)) :
eventually (λ x, P (f x)) F:= H
-- other properties
theorem mapfilter_id (F : filter Y) : mapfilter (λ x, x) F = F := filter.eq rfl
theorem mapfilter_comp (F : filter X) (g : Y → Z) (f : X → Y) :
mapfilter (g ∘ f) F = mapfilter g (mapfilter f F) := rfl
theorem mapfilter_le_mapfilter {F F' : filter X} (f : X → Y) (H : F ≤ F') :
mapfilter f F ≤ mapfilter f F' :=
begin
rewrite filter.le_iff, intro s, rewrite *mem_mapfilter_iff,
apply filter.subset_of_le H
end
-- alternative proof: take s, assume H', by unfold mem; exact H H'
theorem mapfilter_bot (f : X → Y) : mapfilter f ⊥ = ⊥ :=
filter.eq (by rewrite *filter.bot_eq)
theorem mapfilter_sup (f : X → Y) (F F' : filter X) :
mapfilter f (sup F F') = sup (mapfilter f F) (mapfilter f F') :=
filter.eq (ext (take x, !iff.refl))
theorem mapfilter_inf (f : X → Y) (F F' : filter X) :
mapfilter f (inf F F') ≤ inf (mapfilter f F) (mapfilter f F') :=
le_inf (mapfilter_le_mapfilter f !inf_le_left) (mapfilter_le_mapfilter f !inf_le_right)
/- tendsto -/
definition tendsto (f : X → Y) (F₂ : filter Y) (F₁ : filter X) : Prop := mapfilter f F₁ ≤ F₂
theorem tendsto_of_forall_eventually {F₂ : filter Y} {F₁ : filter X} {f : X → Y}
(H : ∀ P, eventually P F₂ → eventually P (mapfilter f F₁)) :
tendsto f F₂ F₁ := H
theorem eventually_mapfilter_of_tendsto {F₂ : filter Y} {F₁ : filter X} {f : X → Y}
{P : Y → Prop} (H : tendsto f F₂ F₁) (H' : eventually P F₂) :
eventually P (mapfilter f F₁) := H H'
theorem tendsto_iff (F₂ : filter Y) (F₁ : filter X) (f : X → Y) :
tendsto f F₂ F₁ ↔ (∀ P, eventually P F₂ → eventually P (mapfilter f F₁)) :=
iff.refl _
theorem tendsto_iff' (F₂ : filter Y) (F₁ : filter X) (f : X → Y) :
tendsto f F₂ F₁ ↔ (∀ P, eventually P F₂ → eventually (λ x, P (f x)) F₁) :=
iff.refl _
theorem tendsto_comp {f : X → Y} {g : Y → Z} {F₁ : filter X} {F₂ : filter Y} {F₃ : filter Z}
(Hf : tendsto f F₂ F₁) (Hg : tendsto g F₃ F₂) : tendsto (g ∘ f) F₃ F₁ :=
tendsto_of_forall_eventually
(take P, suppose eventually P F₃,
have eventually P (mapfilter g F₂), from Hg this,
have eventually (λ y, P (g y)) F₂, from eventually_of_eventually_mapfilter this,
have eventually (λ y, P (g y)) (mapfilter f F₁), from Hf this,
show eventually P (mapfilter g (mapfilter f F₁)), from eventually_mapfilter this)
theorem tendsto_of_ge {f : X → Y} {F₂ F₂' : filter Y} {F₁ : filter X}
(HF₂ : F₂' ≥ F₂) (H : tendsto f F₂ F₁) : tendsto f F₂' F₁ :=
take P, suppose eventually P F₂',
have eventually P F₂, from eventually_of_le this HF₂,
show eventually P (mapfilter f F₁), from H this
theorem tendsto_of_le {f : X → Y} {F₂ : filter Y} {F₁ F₁' : filter X}
(HF₁ : F₁' ≤ F₁) (H : tendsto f F₂ F₁) : tendsto f F₂ F₁' :=
take P, suppose eventually P F₂,
have eventually P (mapfilter f F₁), from H this,
show eventually P (mapfilter f F₁'), from eventually_of_le this (mapfilter_le_mapfilter _ HF₁)
theorem tendsto_of_ge_of_le {f : X → Y} {F₂ F₂' : filter Y} {F₁ F₁' : filter X}
(HF₂ : F₂' ≥ F₂) (HF₁ : F₁' ≤ F₁) (H : tendsto f F₂ F₁) : tendsto f F₂' F₁' :=
tendsto_of_ge HF₂ (tendsto_of_le HF₁ H)
theorem tendsto_id {F : filter X} : tendsto (λx, x) F F :=
take P, assume H, H
theorem tendsto_inf_left {f : X → Y} {F : filter Y} (F₁ : filter X) {F₂ : filter X}
(H : tendsto f F F₁) :
tendsto f F (inf F₁ F₂) :=
tendsto_of_le !inf_le_left H
theorem tendsto_inf_right {f : X → Y} {F : filter Y} {F₁ : filter X} (F₂ : filter X)
(H : tendsto f F F₂) :
tendsto f F (inf F₁ F₂) :=
tendsto_of_le !inf_le_right H
theorem tendsto_sup_left {f : X → Y} (F₁ : filter Y) {F₂ : filter Y} {F : filter X}
(H : tendsto f F₁ F) :
tendsto f (sup F₁ F₂) F :=
tendsto_of_ge !le_sup_left H
theorem tendsto_sup_right {f : X → Y} {F₁ : filter Y} (F₂ : filter Y) {F : filter X}
(H : tendsto f F₂ F) :
tendsto f (sup F₁ F₂) F :=
tendsto_of_ge !le_sup_right H
theorem tendsto_comp_iff_tendsto_mapfilter (f : X → Y) (g : Y → Z) (F₁ : filter X) (F₂ : filter Z) :
tendsto (g ∘ f) F₂ F₁ ↔ tendsto g F₂ (mapfilter f F₁) :=
!iff.refl
/- at_infty -/
section linorderX
variable [linear_strong_order_pair X]
definition at_infty : filter X := Inf ((λ x : X, principal '[x, ∞)) ' univ)
notation `[at ` `∞]` := at_infty
private lemma principal_Ici_le_principal_Ici {x₁ x₂ : X} (H : x₁ ≤ x₂) :
(principal '[x₂, ∞)) ≤ (principal '[x₁, ∞)) :=
principal_le_principal (show '[x₂, ∞) ⊆ '[x₁, ∞), from λ y Hy, le.trans H Hy)
theorem eventually_at_infty {P : X → Prop} {x : X} (H : ∀ y, y ≥ x → P y) :
eventually P at_infty :=
have H' : eventually P (principal '[x, ∞)), from eventually_principal H,
have principal '[x, ∞) ∈ (λ x : X, principal '[x, ∞)) ' univ, from mem_image_of_mem _ !mem_univ,
eventually_Inf this H'
theorem exists_forall_ge_imp_of_eventually_at_infty {P : X → Prop} [inhabited X]
(H : eventually P at_infty) :
∃ x, ∀ y, y ≥ x → P y :=
let S := (λ x : X, principal '[x, ∞)) ' univ in
have H' : ∀₀ F₁ ∈ S, ∀₀ F₂ ∈ S, ∃₀ F ∈ S, F ≤ F₁ ⊓ F₂,
proof
take F₁, suppose F₁ ∈ S, take F₂, suppose F₂ ∈ S,
obtain x₁ [x₁univ F₁eq], from `F₁ ∈ S`,
obtain x₂ [x₂univ F₂eq], from `F₂ ∈ S`,
or.elim (le_or_gt x₁ x₂)
(suppose x₁ ≤ x₂,
have F₂ ≤ F₁, by rewrite [-F₁eq, -F₂eq]; exact principal_Ici_le_principal_Ici this,
exists.intro F₂ (and.intro `F₂ ∈ S` (le_inf this !le.refl)))
(suppose x₂ < x₁,
have x₂ ≤ x₁, from le_of_lt this,
have F₁ ≤ F₂, by rewrite [-F₁eq, -F₂eq]; exact principal_Ici_le_principal_Ici this,
exists.intro F₁ (and.intro `F₁ ∈ S` (le_inf !le.refl this)))
qed,
have principal '[arbitrary X, ∞) ∈ S,
from mem_image_of_mem _ !mem_univ,
have ∃₀ F ∈ S, eventually P F,
from exists_eventually_of_eventually_Inf this H H',
obtain F [FS ePF], from this,
obtain x [xuniv Feq], from FS,
have P ⊇ '[x, ∞), from subset_of_eventually_principal (eq.subst (eq.symm Feq) ePF),
exists.intro x this
theorem eventually_at_infty_iff (P : X → Prop) [inhabited X] :
eventually P at_infty ↔ ∃ x, ∀ y, y ≥ x → P y :=
iff.intro exists_forall_ge_imp_of_eventually_at_infty
(assume H, obtain x Hx, from H, eventually_at_infty Hx)
end linorderX
/- at_ninfty -/
section linorderX
variable [linear_strong_order_pair X]
definition at_ninfty : filter X := Inf ((λ x : X, principal '(-∞, x]) ' univ)
notation `[at ` `-∞]` := at_ninfty
private lemma principal_Iic_le_principal_Iic {x₁ x₂ : X} (H : x₁ ≤ x₂) :
(principal '(-∞, x₁]) ≤ (principal '(-∞, x₂]) :=
principal_le_principal (show '(-∞, x₁] ⊆ '(-∞, x₂], from λ y Hy, le.trans Hy H)
theorem eventually_at_ninfty {P : X → Prop} {x : X} (H : ∀ y, y ≤ x → P y) :
eventually P at_ninfty :=
have H' : eventually P (principal '(-∞, x]), from eventually_principal H,
have principal '(-∞, x] ∈ (λ x : X, principal '(-∞, x]) ' univ, from mem_image_of_mem _ !mem_univ,
eventually_Inf this H'
theorem exists_forall_le_imp_of_eventually_at_ninfty {P : X → Prop} [inhabited X]
(H : eventually P at_ninfty) :
∃ x, ∀ y, y ≤ x → P y :=
let S := (λ x : X, principal '(-∞, x]) ' univ in
have H' : ∀₀ F₁ ∈ S, ∀₀ F₂ ∈ S, ∃₀ F ∈ S, F ≤ F₁ ⊓ F₂,
proof
take F₁, suppose F₁ ∈ S, take F₂, suppose F₂ ∈ S,
obtain x₁ [x₁univ F₁eq], from `F₁ ∈ S`,
obtain x₂ [x₂univ F₂eq], from `F₂ ∈ S`,
or.elim (le_or_gt x₁ x₂)
(suppose x₁ ≤ x₂,
have F₁ ≤ F₂, by rewrite [-F₁eq, -F₂eq]; exact principal_Iic_le_principal_Iic this,
exists.intro F₁ (and.intro `F₁ ∈ S` (le_inf !le.refl this)))
(suppose x₂ < x₁,
have x₂ ≤ x₁, from le_of_lt this,
have F₂ ≤ F₁, by rewrite [-F₁eq, -F₂eq]; exact principal_Iic_le_principal_Iic this,
exists.intro F₂ (and.intro `F₂ ∈ S` (le_inf this !le.refl)))
qed,
have principal '(-∞, arbitrary X] ∈ S,
from mem_image_of_mem _ !mem_univ,
have ∃₀ F ∈ S, eventually P F,
from exists_eventually_of_eventually_Inf this H H',
obtain F [FS ePF], from this,
obtain x [xuniv Feq], from FS,
have P ⊇ '(-∞, x], from subset_of_eventually_principal (eq.subst (eq.symm Feq) ePF),
exists.intro x this
theorem eventually_at_ninfty_iff (P : X → Prop) [inhabited X] :
eventually P at_ninfty ↔ ∃ x, ∀ y, y ≤ x → P y :=
iff.intro exists_forall_le_imp_of_eventually_at_ninfty
(assume H, obtain x Hx, from H, eventually_at_ninfty Hx)
end linorderX
/- approaches infty -/
section approaches_infty
variable [linear_strong_order_pair Y]
abbreviation approaches_infty (f : X → Y) := tendsto f at_infty
notation f ` ⟶ ` ∞ := tendsto f at_infty
notation f ` ⟶ ` -∞ := tendsto f at_ninfty
theorem approaches_infty_intro [inhabited Y] {f : X → Y} {F : filter X}
(H : ∀ y, eventually (λ x, f x ≥ y) F) :
(f ⟶ ∞) F :=
tendsto_of_forall_eventually
(take P,
suppose eventually P at_infty,
obtain z (Hz : ∀ y, y ≥ z → P y), from exists_forall_ge_imp_of_eventually_at_infty this,
have ∀ x, f x ≥ z → P (f x), from take x, Hz (f x),
have eventually (λ x, P (f x)) F, from eventually_mono (H z) this,
show eventually P (mapfilter f F), from eventually_mapfilter this)
theorem approaches_infty_elim {f : X → Y} {F : filter X} (H : (f ⟶ ∞) F) (y : Y) :
eventually (λ x, f x ≥ y) F :=
have eventually (λ x, x ≥ y) at_infty, from eventually_at_infty (take x, suppose x ≥ y, this),
have eventually (λ x, x ≥ y) (mapfilter f F), from H this,
show eventually (λ x, f x ≥ y) F, from eventually_of_eventually_mapfilter this
theorem approaches_infty_iff [inhabited Y] (f : X → Y) (F : filter X) :
(f ⟶ ∞) F ↔ ∀ y, eventually (λ x, f x ≥ y) F :=
iff.intro approaches_infty_elim approaches_infty_intro
theorem approaches_infty_of_eventually_ge [inhabited Y] {f g : X → Y} {F : filter X}
(H : eventually (λ x, f x ≥ g x) F) (H' : (g ⟶ ∞) F) :
(f ⟶ ∞) F :=
approaches_infty_intro (take y,
have eventually (λ x, g x ≥ y) F, from approaches_infty_elim H' y,
show eventually (λ x, f x ≥ y) F, from
eventually_mpr H (eventually_mpr this (eventually_of_forall _ (take x, le.trans))))
end approaches_infty
/- approaches ninfty -/
section approaches_ninfty
variable [linear_strong_order_pair Y]
abbreviation approaches_ninfty (f : X → Y) := tendsto f at_ninfty
notation f ` ⟶ ` -∞ := tendsto f at_ninfty
theorem approaches_ninfty_intro [inhabited Y] {f : X → Y} {F : filter X}
(H : ∀ y, eventually (λ x, f x ≤ y) F) :
(f ⟶ -∞) F :=
tendsto_of_forall_eventually
(take P,
suppose eventually P at_ninfty,
obtain z (Hz : ∀ y, y ≤ z → P y), from exists_forall_le_imp_of_eventually_at_ninfty this,
have ∀ x, f x ≤ z → P (f x), from take x, Hz (f x),
have eventually (λ x, P (f x)) F, from eventually_mono (H z) this,
show eventually P (mapfilter f F), from eventually_mapfilter this)
theorem approaches_ninfty_elim {f : X → Y} {F : filter X} (H : (f ⟶ -∞) F) (y : Y) :
eventually (λ x, f x ≤ y) F :=
have eventually (λ x, x ≤ y) at_ninfty, from eventually_at_ninfty (take x, suppose x ≤ y, this),
have eventually (λ x, x ≤ y) (mapfilter f F), from H this,
show eventually (λ x, f x ≤ y) F, from eventually_of_eventually_mapfilter this
theorem approaches_ninfty_iff [inhabited Y] (f : X → Y) (F : filter X) :
(f ⟶ -∞) F ↔ ∀ y, eventually (λ x, f x ≤ y) F :=
iff.intro approaches_ninfty_elim approaches_ninfty_intro
theorem approaches_ninfty_of_eventually_le [inhabited Y] {f g : X → Y} {F : filter X}
(H : eventually (λ x, f x ≤ g x) F) (H' : (g ⟶ -∞) F) :
(f ⟶ -∞) F :=
approaches_ninfty_intro (take y,
have eventually (λ x, g x ≤ y) F, from approaches_ninfty_elim H' y,
show eventually (λ x, f x ≤ y) F, from
eventually_mpr H (eventually_mpr this (eventually_of_forall _
(take x H₁ H₂, le.trans H₂ H₁))))
end approaches_ninfty
end topology

View file

@ -1,7 +1,7 @@
theories.topology
=================
* [filterlim](filterlim.lean) : a general theory of limits, based on filters
* [approaches](approaches.lean) : a general theory of limits, based on filters
* [basic](basic.lean) : open and closed sets, separation axioms, and generated topologies
* [order_topology](order_topology.lean)
* [continuous](continuous.lean) : continuous functions