- #1
jostpuur
- 2,116
- 19
Let
[tex]
f:]a,b[\to\mathbb{R}
[/tex]
be a continuously differentiable function, where [itex]]a,b[\subset\mathbb{R}[/itex] is some interval, and define
[tex]
f_n:]a,b[\to\mathbb{R},\quad
f_n(x)=\left\{\begin{array}{ll}
&\frac{f(x+1/n)-f(x)}{1/n},\quad x\in ]a,b-1/n[\\
&\textrm{something continuous},\quad x\in [b-1/n, b[\\
\end{array}\right.
[/tex]
Clearly we have a point wise limit [itex]f_n(x)\to f'(x)[/itex] as [itex]n\to\infty[/itex], but how common it is that this convergence is uniform? Is there some well known theorem that says that the convergence is uniform under some assumptions?
[tex]
f:]a,b[\to\mathbb{R}
[/tex]
be a continuously differentiable function, where [itex]]a,b[\subset\mathbb{R}[/itex] is some interval, and define
[tex]
f_n:]a,b[\to\mathbb{R},\quad
f_n(x)=\left\{\begin{array}{ll}
&\frac{f(x+1/n)-f(x)}{1/n},\quad x\in ]a,b-1/n[\\
&\textrm{something continuous},\quad x\in [b-1/n, b[\\
\end{array}\right.
[/tex]
Clearly we have a point wise limit [itex]f_n(x)\to f'(x)[/itex] as [itex]n\to\infty[/itex], but how common it is that this convergence is uniform? Is there some well known theorem that says that the convergence is uniform under some assumptions?
Last edited: