Notes:Ell^p(C) is complete for p between one and positive infinity inclusive

From Maths
Jump to: navigation, search

Statement

p[1,+]¯R the space p(C) is complete.

Proof

Leammas

Lemma 1:

  • Let (xn)nNp(C) be given
    • Suppose (xn)nN is a Cauchy sequence
      • Write them out as follows:
        x1=(xm1)mN=x11, x21, , xk1, x2=(xm2)mN=x12, x22, , xk2, 
      • Now consider the kth column of this table, this gives us the sequence: (xkm)mN
        • We claim that (xkm)mN is a Cauchy sequence itself

Lemma 2:

Let (X,d) be a metric space, then AP(X) is bounded in X if and only if:

  • xXCR>0aA[d(a,x)C]

Proof body

  • Let p[1,+]¯R be given.
    • Let (xn)nNp(C) be a Cauchy sequence in p(C)
      • By lemma 1, we obtain, for each kN the sequence: (xkm)mN and we know this is Cauchy
      • As C is a complete metric space -
        TODO: link to this
        - we see that (xkm)m converges
        • For kN define: tk:=lim which we have just established exists
          • Define \mathbf{t}:\eq(t_1,\ldots,t_k,\ldots) - a sequence
            • We claim that:
              1. \mathbf{t}\in\ell^p(\mathbb{C})
              2. \lim_{n\rightarrow\infty}(\mathbf{x}_n)\eq\mathbf{t}
            • Proof:
              1. That \mathbf{t}\in\ell^p(\mathbb{C})
                • Let \epsilon>0 be given
                  • Since (\mathbf{x}_n)_n is Cauchy there exists N'\in\mathbb{N} such that \forall n,m\in\mathbb{N}[n>m\ge N\implies d(\mathbf{x}_m,\mathbf{x}_n)<\epsilon] Caveat:Or equivalent definition of Cauchy sequence - I initially used (m\ge N\wedge n\ge N)\implies d(\ldots
                  • Warning:"Also (\mathbf{x}_n)_n is bounded - that's all I've written.
                    • Thus, by lemma 2, \exists R\in\mathbb{R}_{>0}\forall n\in\mathbb{N}[d(x_n,(0,\ldots,0,\ldots))<R]
                      • We now show if p\eq +\infty that \mathbf{t}\in\ell^\infty(\mathbb{C})
                        • In this case we see that d(\mathbf{x}_n,\mathbf{0})<R\iff\left[\mathop{\text{Sup} }_{i\in\mathbb{N} }\left(\vert x^i_n\vert\right)<R\right]
                          • Hence by definition of \mathbf{t} 's elements:
                            • \forall i\in\mathbb{N}\left[\vert t_i\vert\eq \lim_{n\rightarrow\infty}\left(\vert x_n^i\vert\right)\le R\right]
                              • "Thus \mathbf{t}\in\ell^\infty(\mathbb{C})" - all I've written.
              2. We now wish to show limit bit.
                • Notice that d(\mathbf{x}_n,\mathbf{x}_m)<\epsilon\iff \big\vert x_n^i-x_m^i\vert<\epsilon
                  • Fix m\ge N, then let n\rightarrow\infty to obtain: \forall i\in\mathbb{N}[\vert t_i-x_m^i\vert\le\epsilon]
                    • Since m\ge N was arbitrary we see: \forall m\in\mathbb{N}\left[m\ge N\implies \mathop{\text{Sup} }_{i\in\mathbb{N} }(\vert t_i-x_m^i\vert)\le\epsilon\right]
                      • Since \epsilon>0 was arbitrary it follows that (\mathbf{x}_n) converges in \ell^\infty(\mathbb{C}) to \mathbf{t} \in\ell^\infty(\mathbb{C})