Definition 1.16 (Subsequences): Let (X,d) be a metric space, let (x^(n))_{n=m}^infty be a
sequence in X, and let (n_j) be a strictly increasing sequence of integers with n_j >= m
(equivalently, m <= n_1 < n_2 < n_3 < ...). The sequence (x^(n_j))_{j=1}^infty is called a
subsequence of (x^(n))_{n=m}^infty.
Equations
Instances For
Lemma 1.2: Let (X,d) be a metric space, (x^{(n)}) a sequence in X, and x0 ∈ X. If
x^{(n)} → x0 as n → ∞, then for every strictly increasing sequence of indices (n_j), the
subsequence (x^{(n_j)}) converges to x0.
Definition 1.17 (Limit point of a sequence): Let (X,d) be a metric space, let m be a
starting index, and let (x^(n))_{n=m}^infty be a sequence in X. A point L ∈ X is called a
limit point of (x^(n))_{n=m}^infty if for every N ≥ m and every ε > 0 there exists an integer
n ≥ N such that d(x^(n), L) ≤ ε.
Equations
- IsLimitPointSeqFrom x m L = ∀ N ≥ m, ∀ ε > 0, ∃ n ≥ N, dist (x n) L ≤ ε
Instances For
A point is a limit point of a sequence from index m if every tail gets within any ε > 0.
Definition 1.18 (Cauchy sequence): Let (X,d) be a metric space, let m ∈ ℕ, and let
(x^(n))_{n=m}^∞ be a sequence in X. The sequence is called a Cauchy sequence if for every
ε > 0 there exists an integer N ≥ m such that d(x^(j), x^(k)) < ε for all j, k ≥ N.
Equations
- IsCauchySeqFrom x m = ∀ ε > 0, ∃ N ≥ m, ∀ j ≥ N, ∀ k ≥ N, dist (x j) (x k) < ε
Instances For
A Cauchy sequence from index m gives a CauchySeq for the shifted tail.
A CauchySeq of the shifted tail yields a Cauchy sequence from index m.
A sequence is Cauchy from index m iff its shifted tail is a CauchySeq.
Helper for Proposition 1.19: a squeeze argument for the bound 1/(k+1).
Proposition 1.19: Let (X,d) be a metric space, let (x^(n))_{n=m}^infty be a sequence in
X, and let L ∈ X. The following statements are equivalent: (i) for every ε > 0 and every
N ≥ m, there exists n ≥ N such that d(x^(n), L) < ε; (ii) there exists a strictly increasing
sequence of integers (n_j)_{j=1}^∞ with n_1 ≥ m such that d(x^(n_j), L) → 0 as j → ∞.
Lemma 1.3 (Convergent sequences are Cauchy): Let (X,d) be a metric space and let
(x^(n))_{n=m}^∞ be a sequence in X such that x^(n) → x0 ∈ X as n → ∞. Then
(x^(n))_{n=m}^∞ is a Cauchy sequence, i.e., for every ε > 0 there exists N ≥ m such that for
all n, n' ≥ N one has d(x^(n), x^(n')) < ε.
A convergent tail of a sequence is Cauchy from its starting index.
Helper for Lemma 1.4: a strictly increasing subsequence is eventually beyond any index.
Helper for Lemma 1.4: pick a subsequence index beyond N that is ε-close to x0.
Lemma 1.4: Let (X,d) be a metric space and let (x^{(n)})_{n=m}^∞ be a Cauchy sequence in
X. Suppose there exists a subsequence (x^{(n_j)}) and a point x0 ∈ X such that
x^{(n_j)} → x0 as j → ∞. Then x^{(n)} → x0 as n → ∞.
Helper for Lemma 1.5: a strictly increasing index map dominates any lower bound.
Lemma 1.5 (Subsequence of Cauchy sequence is Cauchy): Let (X,d) be a metric space and let
(x^(n))_{n=m}^∞ be a Cauchy sequence in X. Then every subsequence (x^(n_j))_{j=1}^∞ of
(x^(n))_{n=m}^∞ is also a Cauchy sequence.
Definition 1.19 (Complete metric space): A metric space (X,d) is complete if every Cauchy
sequence converges to a point of X, i.e., for every Cauchy sequence there exists x ∈ X such
that d(x_n, x) → 0 as n → ∞.
Equations
- IsCompleteMetricSpace X = ∀ (x : ℕ → X), CauchySeq x → ∃ (L : X), Filter.Tendsto x Filter.atTop (nhds L)
Instances For
The book's Cauchy-sequence definition of completeness is equivalent to CompleteSpace.
Proposition 1.20: Let (X,d) be a metric space and let Y ⊆ X be equipped with the
subspace metric d|_{Y×Y}. (a) If (Y, d|_{Y×Y}) is complete, then Y is closed in X. (b) If
(X,d) is complete and Y is closed in X, then (Y, d|_{Y×Y}) is complete.
Proposition 1.21: Let (X,d) be a metric space, let m ∈ ℕ, and let (x^(n))_{n=m}^∞ be a
sequence in X. Let L ∈ X. Assume that L is a limit point of the sequence, i.e., for every
ε > 0 and every N ≥ m there exists n ≥ N such that d(x^(n), L) < ε. Then L is an
adherent point of the set A := {x^(n) : n ≥ m}, i.e., for every ε > 0 there exists a ∈ A
such that d(a, L) < ε.
A limit point of a tail of a sequence is an adherent point of the tail set.
Proposition 1.22: Let (X,d) be a metric space, and let (x_n)_{n∈ℕ} be a Cauchy sequence
in X. If x_n → x and x_n → y for some x,y ∈ X, then x = y. Equivalently, a Cauchy
sequence in a metric space has at most one limit.
A Cauchy sequence in a metric space has at most one limit.
Definition 1.20 (Completion via formal limits): Let (X,d) be a metric space. (1) For every
Cauchy sequence (x_n)_{n=1}^∞ in X, introduce a formal symbol LIM_{n→∞} x_n. (2) Define an
equivalence relation ~ on such formal symbols by LIM x_n ~ LIM y_n iff
lim_{n→∞} d(x_n,y_n) = 0. (3) Define X̄ to be the set of equivalence classes modulo ~. (4)
Define d_{X̄} : X̄ × X̄ → [0,∞) by d_{X̄}(LIM x_n, LIM y_n) := lim_{n→∞} d(x_n,y_n), where
representatives are chosen from the corresponding equivalence classes.
Instances For
The relation identifying two formal limits when their pointwise distance tends to 0.
Equations
- FormalLimitRel X x y = Filter.Tendsto (fun (n : ℕ) => dist (↑x n) (↑y n)) Filter.atTop (nhds 0)
Instances For
Helper for Proposition 1.23: reflexivity of the formal-limit relation.
Helper for Proposition 1.23: symmetry of the formal-limit relation.
Helper for Proposition 1.23: transitivity of the formal-limit relation.
Proposition 1.23: Let (X,d) be a metric space, and let C be the set of all Cauchy
sequences in X. For Cauchy sequences (x_n) and (y_n) define a relation ~ on the set of
formal limits {LIM_{n→∞} x_n : (x_n) ∈ C} by LIM x_n ~ LIM y_n iff
lim_{n→∞} d(x_n,y_n) = 0. Then ~ is an equivalence relation.
Helper for Proposition 1.24: the setoid induced by the formal-limit relation.
Equations
- FormalLimitSetoid X = { r := FormalLimitRel X, iseqv := ⋯ }
Instances For
Helper for Proposition 1.24: the completion via formal limits as a quotient of formal symbols.
Equations
Instances For
Helper for Proposition 1.24: the pointwise distance between two Cauchy representatives is itself a Cauchy sequence.
Helper for Proposition 1.24: define the representative-level distance via limUnder.
Equations
- helperForProposition_1_24_formalLimitDist X x y = limUnder Filter.atTop fun (n : ℕ) => dist (↑x n) (↑y n)
Instances For
Helper for Proposition 1.24: the representative-level distance is the limit of pointwise distances.
Helper for Proposition 1.24: the representative-level distance respects the formal-limit equivalence relation.
Helper for Proposition 1.24: the distance on the formal completion, defined by limits of distances.
Equations
- completionViaFormalLimitsDist X qx qy = Quotient.liftOn₂ qx qy (fun (x y : FormalLimit X) => helperForProposition_1_24_formalLimitDist X x y) ⋯
Instances For
Helper for Proposition 1.24: the coerced sequence of a formal limit is Cauchy in the completion.
Helper for Proposition 1.24: send a formal limit to the limit of its coerced sequence.
Equations
- helperForProposition_1_24_formalLimit_to_uniformCompletion X x = limUnder Filter.atTop fun (n : ℕ) => ↑(↑x n)
Instances For
Helper for Proposition 1.24: the coerced sequence converges to its chosen completion point.
Helper for Proposition 1.24: equivalent formal limits map to the same completion point.
Helper for Proposition 1.24: the map from the formal completion to the uniform completion.
Equations
Instances For
Helper for Proposition 1.24: powers of 1/2 tend to 0.
Helper for Proposition 1.24: the map to the uniform completion is surjective.
Helper for Proposition 1.24: the map to the uniform completion is injective.
Helper for Proposition 1.24: the formal-limit completion is (noncanonically) equivalent to the uniform completion.
Helper for Proposition 1.24: dist vanishes on identical classes.
Helper for Proposition 1.24: the formal-limit distance is symmetric.
Helper for Proposition 1.24: the formal-limit distance satisfies the triangle inequality.
Helper for Proposition 1.24: distance zero gives equality in the quotient.
Helper for Proposition 1.24: the edist field is ENNReal.ofReal of the distance.
Helper for Proposition 1.24: the pseudo-metric structure on the formal completion.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Helper for Proposition 1.24: the formal-limit completion carries a metric-space structure.
Equations
- helperForProposition_1_24_completionMetricSpace X = { toPseudoMetricSpace := helperForProposition_1_24_completionPseudoMetricSpace X, eq_of_dist_eq_zero := ⋯ }
Instances For
Proposition 1.24: Let (X,d) be a metric space. Define X̄ as the set of equivalence classes
of Cauchy sequences in X, where two Cauchy sequences (x_n) and (y_n) are equivalent if
d(x_n,y_n) → 0 as n → ∞. For classes [(x_n)], [(y_n)] ∈ X̄, define
d_{X̄}([(x_n)],[(y_n)]) := lim_{n→∞} d(x_n,y_n). Then d_{X̄} is well-defined and
(X̄, d_{X̄}) is a metric space.
Helper for Proposition 1.24: the formal-limit distance does not depend on the chosen Cauchy representatives.
Helper for Proposition 1.24: the formal-limit completion carries a metric-space structure via the formal-limit distance.
Helper for Theorem 1.4: choose representatives for a sequence in the quotient.
Helper for Theorem 1.4: reindexing a Cauchy representative stays Cauchy.
Helper for Theorem 1.4: a formal limit is equivalent to any strictly monotone reindexing.
Helper for Theorem 1.4: if the formal-limit distance is below r, then pointwise distances are eventually below r.
Helper for Theorem 1.4: extract a geometric subsequence of a Cauchy sequence in the completion.
Helper for Theorem 1.4: build a monotone diagonal index with pointwise control.
Helper for Theorem 1.4: the diagonal sequence is Cauchy.
Helper for Theorem 1.4: a Cauchy sequence with a convergent subsequence converges.