16 CHAPTER 1. BASIC NOTIONS

That is, it consists of ordered lists of p numbers from F. These will be denoted as x, boldfaced. For now, ∥x∥

∞≡ max{|xk| : k = 1,2, ..., p}. Thus, to say that xk → x will mean

that limk→∞ ∥xk−x∥∞= 0 which happens if and only if the entries of xk converge to the

corresponding entries of x. This is called a norm and more will be said about these later.The following is important.

Axioms of a Norm

∥x∥ ≥ 0 and ∥x∥= 0 if and only if x = 0, (each xk = 0) (1.1)

If α ∈ F, then ∥αx∥= |α|∥x∥ (1.2)

∥x+y∥ ≤ ∥x∥+∥y∥ (1.3)

Only the last property is not obvious. However,

∥x+y∥ ≡ max{|xk + yk| : k ≤ p} ≤max{|xk|+ |yk| : k ≤ p}≤ max{|xk| : k ≤ p}+max{|yk| : k ≤ p} ≡ ∥x∥+∥y∥ (1.4)

Recall that F is complete. See my book Analysis of Functions of One Variable, for example.It follows easily that Fp is also complete because any Cauchy Sequence in Fp has each entrya Cauchy sequence in F and so it converges. This is in the following proposition.

Definition 1.6.1 {xn} is a Cauchy sequence in Fp means that for all ε > 0 thereexists nε such that if m,n≥ nε , then ∥xn−xm∥< ε . A sequence {xn} is said to converge ifthere exists x such that limn→∞ xn = x.

Proposition 1.6.2 If {xn}∞

n=1 is a Cauchy sequence in Fp, then there exists x∈ Fp suchthat limn→∞ ∥xn−x∥= 0.

Proof: For each k,{

xnk

}is a Cauchy sequence. Thus, there exists xk ∈ F such that

limn→∞ xnk = xk. Therefore, letting x≡ (x1, · · · ,xp) , limn→∞ ∥xn−x∥= 0. ■

Definition 1.6.3 Letting {xn} be a sequence of vectors in Fp, ∑∞k=1 xk is said to

converge if there exists s such that limn→∞ ∑nj=1 x j = s.

The Weierstrass M test is a convenient way to consider convergence of series in Fp.

Proposition 1.6.4 If there exists Mk such that Mk ≥∥∥xk∥∥ , and if ∑k Mk converges, then

so does ∑k xk.

Proof: For m < n, ∥∥∥∥∥ n

∑k=1

xk−m−1

∑k=1

xk

∥∥∥∥∥≤ n

∑k=m

∥∥∥xk∥∥∥≤ ∞

∑k=m

Mk

and if m is large enough, the term on the right is no more than ε by the standard material inCalculus. Therefore, the partial sums are a Cauchy sequence and must converge thanks toProposition 1.6.2. ■

This is not the best norm for Fp however. That will be described next.

16 CHAPTER 1. BASIC NOTIONSThat is, it consists of ordered lists of p numbers from F. These will be denoted as x, boldfaced. For now, ||x||,, = max {|x;,|:k = 1,2,...,p}. Thus, to say that x, — x will meanthat limg +00 ||xz — x||,, = 0 which happens if and only if the entries of x, converge to thecorresponding entries of x. This is called a norm and more will be said about these later.The following is important.Axioms of a Norm||x|| > 0 and ||x|| = 0 if and only if x = 0, (each x, = 0) (1.1)If a € F, then ||ox|| = |e ||x|| (1.2)IIx +y|] < |[x|] + |lyl| (1.3)Only the last property is not obvious. However,|Ixt+y|| = max{|x,+y,|:k < p} < max {|xz|4+ |yg| & < p}< max {|xg| 2k < p} +max {|ye| 2k < p} = |x| +llyll (1.4)Recall that F is complete. See my book Analysis of Functions of One Variable, for example.It follows easily that F? is also complete because any Cauchy Sequence in F? has each entrya Cauchy sequence in F and so it converges. This is in the following proposition.Definition 1.6.1 {x"} is a Cauchy sequence in F? means that for all € > 0 thereexists ng such that if m,n > ne, then \|x" —x"|| < €. A sequence {x"} is said to converge ifthere exists x such that limp-5..X”" = x.Proposition 1.6.2 If {x"};_, is a Cauchy sequence in F”, then there exists x € F? suchthat liMy-+00 ||x" — x|| = 0.Proof: For each k, {xj} is a Cauchy sequence. Thus, there exists x, € F such thatlimy 00%} = xx. Therefore, letting x = (x1,--- .Xp), limp. ||x"” —x|| = 0.Definition 1.6.3 Letting {x"} be a sequence of vectors in F?, Y°_,x* is said toconverge if there exists s such that limy +. yj- |x/=s.The Weierstrass M test is a convenient way to consider convergence of series in F?.Proposition 1.6.4 /f there exists M;, such that M, > ||x*so does x".| , and if YM, converges, thenProof: For m <n,n m—1yx- yxk=1 k=1< Esl Emk=m k=mand if m is large enough, the term on the right is no more than € by the standard material inCalculus. Therefore, the partial sums are a Cauchy sequence and must converge thanks toProposition 1.6.2.This is not the best norm for F? however. That will be described next.