1
$\begingroup$

I am working through two proofs that there exists a Schauder Basis for $C([0,1])$.

One proof defines a basis $(f_n)_{n=0}^{\infty}$ with $$ f_0(x) = 1 \qquad f_1(x) = x $$ for $2^{k-1} < n \le 2^{k}$, where $k \ge 1$, we define $$ f_n(x) = \left\{ \begin{array}{ll} 2^k ( x - (2^{-k}(2n - 2) - 1)) & \mathrm{if} ~ x \in I_n \\ 1 - 2^k ( x - (2^{-k}(2n - 1) - 1)) & \mathrm{if} ~ x \in J_n \\ 0 & \mathrm{otherwise} \end{array} \right. $$ where $$ I_n = [2^{-k}(2n-2), 2^{-k}(2n-1)) \qquad J_n = [2^{-k}(2n-1), 2^{-k}2n). $$ The graphs of these functions form a sequence of "tents" of height one and width $2^{-k+1}$ that sweep across the interval $[0,1]$.

This proof is from this notes page 94.

Another proof i found on the internet goes like this, define the "triangle function" $$ \Delta(x) = \begin{cases} 2x & \mathrm{if } \hspace{2mm} x \in \left[0, \frac{1}{2}\right] \\ \\ 2(1-x) & \mathrm{if } \hspace{2mm} x \in \left(\frac{1}{2},1\right] \\ \\ 0 & \mathrm{otherwise}. \end{cases}$$ Then consider for $n > 0$ $$ \Delta_n(x) = \Delta(2^j x - k) \quad \mathrm{for } \qquad n = 2^j + k, \quad j \ge 0,\quad 0 \le k < 2^j $$ and $\Delta_{-1}(x) = 1, \Delta_0(x) = x$. Then the Sequence $\Delta_{-1}, \Delta_0, \Delta_1, \Delta_2, \ldots$ forms a Schauder basis for the Banach space of continuous functions on $[0,1]$.

This proof i found there and here. Now here's my question. I tried to prove that both, the $\Delta_n$ and the $f_n$'s define the same functions, but i am not able to convert the definitions to each other. Do you know how can i proof that these two functions are essentially the same "tent"-functions?

$\endgroup$
  • $\begingroup$ @Martin Argerami, may be you meant Hamel basis? $\endgroup$ – Norbert Jun 18 '12 at 6:36
  • $\begingroup$ Of course I did. Thanks! $\endgroup$ – Martin Argerami Jun 18 '12 at 14:29
  • $\begingroup$ I am reading the pg 94 of the notes that you provided a link to, and could you explain to me the following? The author says at the end of the example that the "uniform continuity of f implies that the series converges uniformly. But, why is that so? $\endgroup$ – yumiko Dec 21 '16 at 19:11
2
$\begingroup$

Functions $f_n$ specified incorrectly. The correct ones formulas are $$ f_n(x) = \begin{cases} 2^k ( x - (2^{-k}(2n - 2) - 1)) & \text{ if } \quad x \in I_n \\ 1 - 2^k ( x - (2^{-k}(2n - 1) - 1)) & \text{ if } \quad x \in J_n \\ 0 & \text{ otherwise } \end{cases}. $$ where $$ I_n = [2^{-k}(2n-2)-1, 2^{-k}(2n-1)-1) \qquad J_n = [2^{-k}(2n-1)-1, 2^{-k}2n-1). $$ In order to prove that $f_n=\Delta_{n-1}$ rewrite these formulas as $$ f_n(x) = \begin{cases} 2^{j+1} ( x - (2^{-j-1}(2n - 2) - 1)) & \text{ if } \quad 2^{-j-1}(2n-2)-1\leq x<2^{-j-1}(2n-1)-1\\ 1 - 2^{j+1} ( x - (2^{-j-1}(2n - 1) - 1)) & \text{ if } \quad 2^{-j-1}(2n-1)-1\leq x<2^{-j-1}2n-1 \\ 0 & \text{ otherwise } \end{cases}= $$ $$ \begin{cases} 2^{j+1} x - n + 1 + 2^{j+1} & \text{ if } \quad (2n-2)\leq 2^{j+1}(x+1)<(2n-1)\\ 2n-2^{j+1}x-2^{j+1} & \text{ if } \quad 2n-1\leq 2^{j+1}(x+1)<2n \\ 0 & \text{ otherwise } \end{cases}= $$ $$ \begin{cases} 2(2^{j} x - (n - 1 - 2^{j})) & \text{ if } \quad 0\leq 2^{j} x - (n - 1 - 2^{j})<\frac{1}{2}\\ 2-2(2^jx-(n-1-2^j)) & \text{ if } \quad \frac{1}{2}\leq 2^{j} x - (n - 1 - 2^{j})<1 \\ 0 & \text{ otherwise } \end{cases}= $$ $$ \Delta(2^jx-(n-1-2^j))=\Delta(2^jx-k)=\Delta_{n-1}(x) $$ And now we are done.

$\endgroup$
  • $\begingroup$ There's something wrong in your code for $f_n$, as the functions given in the OP definitely are functions on $[0,1]$ while there's something wiggling over $[1,2]$ in your plots. $\endgroup$ – t.b. Jun 21 '12 at 22:05
  • $\begingroup$ Of course because those functions was incorrect. Now I've written correct formulas. $\endgroup$ – Norbert Jun 21 '12 at 22:09
  • $\begingroup$ I see. Indexing problems. +1 for wading through that horrible mess :) $\endgroup$ – t.b. Jun 21 '12 at 22:15
  • $\begingroup$ @t.b. Thanks! ${}{}{}$ $\endgroup$ – Norbert Jun 21 '12 at 22:19

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.