-1
$\begingroup$

Suppose we have the series of functions:

\begin{equation} F(x)=\sum_{n=1}^{\infty} f_n(x) \end{equation}

where convergence is uniform.

Additionally, consider the partial functions of the series:

\begin{equation} F_m (x)=\sum_{n=1}^{m} f_n(x) \end{equation}

Suppose that:

\begin{equation} \int_{0}^{1} \frac{F_m(x)}{x^{3}} ,dx \end{equation}

forms a Cauchy sequence, hence convergent.

Can we then claim that:

\begin{equation} \int_{0}^{1} \frac{F_m(x)}{x^{3}} ,dx \to \int_{0}^{1} \frac{F(x)}{x^{3}} ,dx \end{equation}

as $m \to \infty$?

observation: \begin{equation} \int_{0}^{1} \frac{F(x)}{x^{3}} ,dx \end{equation}

exist

$\endgroup$
1
  • 1
    $\begingroup$ What do you mean by "Obs:"? $\endgroup$ Commented Dec 13, 2023 at 20:54

2 Answers 2

0
$\begingroup$

The answer is no with a simple counterxample: consider $G_n(x)=n\chi_{[0,1/n]}(x)$ and $F_n(x)=x^3G_n(x), 0 \le F_n(x) \le 1/n^2$

(so one can define $f_n=F_n-F_{n-1}$)

Then clearly $F_n \to 0$ uniformly and $\int_0^1G_n(x)dx=1 \to 1$, while $F=0$ so $\int_0^1 F(x)dx/x^3=0$

With a little smoothing one clearly can make everything continuous or differentiable as wished etc

$\endgroup$
4
  • $\begingroup$ but in this case your sequence of functions is unlimited, if I have a limited sequence? $\endgroup$ Commented Dec 13, 2023 at 21:09
  • $\begingroup$ what do you mean limited? $F_n/x^3$ bounded? $\endgroup$
    – Conrad
    Commented Dec 13, 2023 at 21:14
  • $\begingroup$ no, $F_n(x)$ bounded. $\endgroup$ Commented Dec 13, 2023 at 21:15
  • $\begingroup$ $F_n$ is bounded by $1/n^2$ in my counterxample, $F_n/x^3$ is unbounded $\endgroup$
    – Conrad
    Commented Dec 13, 2023 at 21:15
0
$\begingroup$

No.

Indeed, make the substitution $x=u^{-1/2}$. That is, let $G_n(u):=F_n(u^{-1/2})$ and $G(u):=F(u^{-1/2})$, so that $\int_0^1\frac{F_n(x)}{x^3}\,dx=\frac12\,\int_1^\infty G_n$ and $\int_0^1\frac{F(x)}{x^3}\,dx=\frac12\,\int_1^\infty G$. So, we can restate the question as follows:

Let $(G_n)$ be a sequence of functions on $[1,\infty)$ converging uniformly to a function $G$ on $[1,\infty)$ so that $\int_1^\infty G_n$ converges (as $n\to\infty$) and $\int_1^\infty G$ exists. Does it then follow that $\int_1^\infty G_n\to\int_1^\infty G$?

To show that this is not true, let e.g. $$G_n:=\frac1n\,1_{[n,2n]}.$$ Then $G_n\to0=:G$ uniformly but $$\int_1^\infty G_n=1\not\to0=\int_1^\infty G.$$

$\endgroup$
2
  • $\begingroup$ @georgeandrade : This change of the integrands and of the interval of integration (which results in simpler integrands) is obtained by substitution, as is now explained in detail. Please let me know if further details are needed. $\endgroup$ Commented Dec 13, 2023 at 21:21
  • $\begingroup$ @georgeandrade : Do you have a further response to the answers on this page? $\endgroup$ Commented Dec 14, 2023 at 23:41

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .