Limit Laws (Algebraic Limit Theorem)

Algebraic Limit Theorem

Let (an),(bn) be convergent sequences, where an→a,bn→b. Then:

  1. For c∈R, lim(can)=ca.
  2. lim(an+bn)=a+b
  3. lim(anbn)=ab
  4. lim(anbn)=ab given b≠0. (ignore the cases when bn=0 for the inner sequence. At most finitely many bn=0)

Scratch Work:

  1. We want, for some ε>0:
|can−ca|<ε⇒|c||an−a|<ε

Consider the cases. If c=0 then the case is trivial. If c≠0 then:

|an−a|<ε|c|

We know that ∃N such that for all n≥N.

  1. We want, for some ε>0:
|(an+bn)−(a+b)|<ε

Notice then:

|(an−a)+(bn−b)|≤|an−a|⏟<ε2+|bn−b|⏟<ε2
  1. We want, for some ε>0:
|anbn−ab|<ε

Notice that, ignoring the absolute values for a moment:

anbn−ab−abn+abn=(an−a)bn+a(bn−b)

So putting the absolute values back in:

|(anbn)−ab|=|(an−a)bn+a(bn−b)|≤|bn||an−a|⏟<ε2M+|a||bn−b|⏟<ε2(|a|)∨ε2(|a|+1)Triangle Ineq.

where since bn is convergent, then it has a bound M, so then |bn|≤M. Notice further that M>0 since we consider the absolute value of bn.

  1. Say b≠0. Then we want:
1bn→1b

Because then we can apply (3) by using the new sequence 1/bn instead. This means we need to show:

|1bn−1b|=|b−bn||b||bn|

Note that because (bn)→b, then we can make the numerator as small we we like by making n large. On the other hand, for the denominator we want some |bn|≥δ>0. This will lead to a bound on the size of 1|b||bn|.

The trick is to look far enough out into the sequence (bn) so that the terms are closer to b than they are to 0. Consider ε0=|b|2. Because (bn)→b then there exists and N1 such that |bn−b|<|b|2 for all n≥N1. This implies that |bn|>|b|2. So then choose N2 such that N≥N2 implies:

|bn−b|<ε|b|22

and choose N=max(N1,N2).

Proof

  1. Let ε>0. Choose N∈N such that:
    a. If c=0 then we are done (trivial) case.
    b. Since c≠0 then because ε|c|>0 then n≥N⇒|an−a|<ε|c| . Then:
|an−a|<ε|c||c||an−a|<ε|can−ca|<ε
  1. Let ε>0. Since ε2>0 then ∃N1 where n≥N1 implies |an−a|<ε2. Similarly then ∃N2 where n≥N2 implies |bn−b|<ε2. Choose N=max(N1,N2)∈N, since then letting n≥N:
|(an+bn)−(a+b)=|(an−a)+(bn−b)|≤|an−a|+|bn−b|ΔInequality<ε2+ε2=ε
  1. Let ε>0. Since bn→b then ∃M≥0 where |bn|≤M. Notice that:
ε>0⇒ε2M>0

And further:

ε>0⇒ε2(|a|+1)>0

(or choose ε/2(|a|) if |a|≠0. Either way choosing either doesn't affect the later arguments). Then since an→a then ∃N1 where for all n≥N1 then |an−a|<ε2(|a|+1). Similarly since bn→b then ∃N2 where for all n≥N2 when |bn−b|<ε2M.

Now choose N=max(N1,N2). Let n≥N. Then:

|anbn−ab|=|anbn−abn+abn−ab|=|(an−a)bn+a(bn−b)|≤|bn||an−a|+|a||bn−b|<|bn|ε2M+|a|ε2(|a|+1)<ε2⏟|bn|≤M+ε2⏟|a|<ε2(|a|+1)=ε
  1. If we can prove that 1bn→1b then we can apply (3) and we are done (assuming b≠0). Let ε>0. First note:
|1bn−1b|=|b−bn||b||bn|

Now because (bn)→b then |bn−b|<ε|b|22 since it's >0. There's some N2 such that n≥N2 allows for this inequality to hold.

Now notice also that because (bn)→b then notice for a specific ε0=|b|2 that then ∃N1 where |bn−b|<|b|2 for all n≥N1. Then this implies that |bn|>|b|2.

Thus choose N=max(N1,N2) so then for all n≥N:

|1bn−1b|=|b−bn|1|b||bn|<ε|b|22⋅1|b||b|2=ε

☐