Prove this inequality (geometric series and optimisation)

64 Views Asked by At

Show that

$$\sum_{n = 0}^{+\infty} \frac{\ln(n^3+1)}{n^3+1} \frac{1}{4^n} \leq \frac{\ln(2)}{6}$$

I don't know where I am wrong, but I don't get the desired result. The hint was "use the function $\frac{\ln(x^3+1)}{x^3+1}$ in the reasoning.

attempts

So I thought that if $f(x) = \frac{\ln(x^3+1)}{x^3+1}$ is limited, then I could say

$$\sum_{n = 0}^{+\infty} \frac{\ln(n^3+1)}{n^3+1}\frac{1}{4^n} \leq C \sum_{n = 0}^{+\infty} \frac{1}{4^n}$$

Which I can then sum as a Geometric Series.

So $f(x) = \frac{\ln(x^3+1)}{x^3+1}$, and by studying the derivative I got that $f(x)$ has a maximum at $x = \sqrt[3]{e-1}$ hence $$\text{max}(f(x)) = f(\sqrt[3]{e-1}) = \frac{1}{e}$$

Yet by doing so I obtain

$$\sum_{n = 0}^{+\infty} \frac{\ln(n^3+1)}{n^3+1}\frac{1}{4^n} \leq \frac{1}{e}\sum_{n = 0}^{+\infty} \frac{1}{4^n} = \frac{4}{3e}$$

But it is surely not true that $\frac{4}{3e} \leq \frac{\ln(2)}{6}$

So where I am wrong?

Thank you!

1

There are 1 best solutions below

0
On

The slightly weaker bound you got was because we are actually only evaluating $f$ at positive integers. Calculus shows that $f(x)$ is decreasing for $x$ larger than the extremum you found, and comparing $f(1)$ and $f(2)$ shows us that $f(1)\geq f(n)$ for all positive integers $n$.

Then your sum can be estimated as $$\sum_{n=1}^\infty f(n)4^{-n}\leq f(1)\sum_{n=1}^\infty 4^{-n}=\frac{\log 2}{6}. $$