0

Problem: Given a continuous function $f(x)\geq 0$ defined on $[a,b]$, let $$ M_n = \left[\int_a^b (f(x))^ndx\right]^{\frac{1}{n}}. $$ Show that $M_n \rightarrow \sup_{x \in [a,b]} f(x)$ as $n \rightarrow \infty$.

My Attempt: I had a hard time trying to find a way to show the equality so I just tried to use squeeze theorem, let me know if this approach works/ if there's a much better way to approach the problem. Thanks in advance!

Because $f(x)$ is continuous on a compact set it is bounded by the min-max theorem; and in particular there are values on the interval for which it attains it's supremum and infimum. So if we let $m = \inf_{x \in [a,b]} f(x)$ and $M = \sup_{x \in [a,b]} f(x)$ then $m \leq f(x) \leq M$. It follows that we can bound $M_n$ as $$ M_n \leq \left[\int_a^bM^ndx\right]^{\frac{1}{n}}. = \left[M^n \int_a^b dx \right]^{\frac{1}{n}} = \left[M^n(b-a)\right]^{\frac{1}{n}} = M(b-a)^\frac{1}{n}. $$ Note also that because $f(x) \geq 0$ we can equivalently bound $f$ as $-M \leq f \leq M$ and moreover $-M^n \leq f^n \leq M^n$. Note that we know that the product of Riemann integrable functions is also riemann integrable; so $f^n \in \mathscr{R}$, and so we can conclude that for any partition $P$ of $[a,b]$ that $$ -M^n(b-a) \leq L(P,f^n) \leq \overline{\int_{a}^{b}} f^ndx = \int_a^b f^ndx. $$ (Here the overlined integral denotes the upper Riemann integral) Note above the inequalities are because $f^n$ is bounded, the rectangle is less than any lower sum which is itself bounded by an upper Riemann integral, and because $f^n \in \mathscr{R}$ the upper integral is the integral. From this it follows that \begin{align*} (M^n(a-b))^\frac{1}{n} &\leq \left[\int_a^b f^n dx\right]^\frac{1}{n} \\ M (a-b)^\frac{1}{n} &\leq M_n. \end{align*} Now note that if we combine the two inequalities we get that $$ M (a-b)^\frac{1}{n} \leq M_n \leq M(b-a)^\frac{1}{n}. $$ and taking $n\rightarrow \infty$ across the inequalities gives that $$ M\leq M_n \leq M, $$ and so we conclude that $\lim_{n \rightarrow \infty} M_n = M = \sup_{x \in [a,b]} f(x)$.

Irving Rabin
  • 2,643
  • not sure if it's valid to to take the limit of a negative number to the $\frac 1 n$, since for every even $n$ you leave the real numbers.... – Alan Apr 23 '22 at 01:44
  • Ah damn I didn't think that through, thanks. I couldn't figure out how to get that bound so I forced it. Any way to salvage the argument? – Irving Rabin Apr 23 '22 at 01:48
  • not that I can think of – Alan Apr 23 '22 at 01:52
  • Any recommendations for a proper approach? – Irving Rabin Apr 23 '22 at 01:52
  • Well, my intuition is that since $x^n$ is increasing on $x>0$ and your functions are increasing, the impact of your integral is more strongly impacted by the area around the max compared to anywhere else. Might want to see if that leads you somewhere? – Alan Apr 23 '22 at 01:54
  • Thanks! I'll see what that looks like. – Irving Rabin Apr 23 '22 at 01:55

0 Answers0