Convergence of Sequences in Metric Spaces
📂MetricSpaceConvergence of Sequences in Metric Spaces
Definitions
If there exists a point p∈X such that the sequence {pn} of points in a metric space (X,d) satisfies the following condition, the sequence {pn} is said to converge to p, and it is denoted by pn→p or n→∞limpn=p.
∀ε>0, ∃N∈N s.t n≥N⟹d(pn,p)<ε
If {pn} does not converge, it diverges. Furthermore, the set of all pns is called the range of {pn}. If the range of {pn} is bounded, the sequence {pn} is said to be bounded.
Theorems
Let {pn} be a sequence in a metric space (X,d).
(a) A necessary and sufficient condition for pn→p is that every neighborhood of p contains all terms of {pn} except for a finite number.
(b) If pn→p and pn→p′, then p=p′.
(c) If {pn} converges, it is bounded.
(d) Given E⊂X. If p is a limit point of E, there exists a sequence {pn} from E that satisfies p=n→∞limpn. Furthermore, if {pn} is a set of distinct points, the converse also holds.
Proofs
(a)
(⟹)
Suppose pn→p. Given any positive number ε>0, let V be a neighborhood of p with radius ε. According to the definition of a neighborhood, the following holds:
d(p,q)<ε⟹q∈V
However, by assumption, there exists a N that satisfies the following condition for the given ε:
∀n≥N, d(pn,p)<ε
Therefore, all points pn, except for a finite number, are included in V.
(⟸)
Assume every neighborhood of p contains all but a finite number of points {pn}. Given any positive number ε>0, let V be a neighborhood of p with radius ε. Then, by assumption, a N exists that satisfies the following condition:
n≥N⟹pn∈V
Since V is a neighborhood of p, the following holds:
∀n≥N,d(pn,p)<ε
Therefore, pn→p.
■
(b)
Given any positive number ε>0. By assumption, there exist two positive numbers N, N′ that satisfy the following condition:
n≥Nn≥N′⟹d(pn,p)<2ε⟹d(pn,p)<2ε
Then, for n≥max(N,N′), the following equation holds:
d(p,p′)≤d(p,pn)+d(pn,p′)<ε
Since ε is any positive number,
d(p,p′)=0
and by the definition of distance, p=p′.
■
(c)
Assume that {pn} converges to p. By assumption, there exists a positive number N that satisfies the following equation:
n≥N⟹d(pn,p)<1
Now, let
r=max{1, d(p1,p), ⋯, d(pN,p)}
Then, for all n,
d(pn,p)≤r
so {pn} is bounded.
■
(d)
(⟹)
Assume E⊂X and that p is a limit point of E. By definition of a limit point, for each n,
d(pn,p)<n1
exists a pn∈E. Now, given any positive number ε>0 and Nε>1 satisfying N, the following holds for n>N:
d(pn,p)<n1<nNε<ε
Therefore, {pn} converges to p.
(⟸)
Assume there exists a sequence {pn} of distinct points from E that satisfies p=n→∞limpn. Then, for all positive numbers ε>0,
n≥N⟹d(pn,p)<ε
exists a N. Let Vε be a neighborhood of p with radius ε. Thus, Vε includes pn∈E(n≥N) which is not p, proving p is a limit point of E.
■