Dear SE Mathematics Community,
I have encountered a little riddle in probability theory, that I would like to put out there.
Question:
Let $I_n=\{\frac{k}{n}; k \in \{1, 2, ..., n\}\}$. Let $X_n:\Omega \to I_n$ the random variable, such that $P(X_n = \frac{k}{n})= \frac{a_nk}{n^2+k^2}$.
- What is the limit of the sequence $(a_n)$?
- Prove that $lim E[X_n]=\frac{4-\pi}{2\log(2)}$.
It seems obvious that we can only prove (2.) if we know the answer to (1.). Yet, I seem to run into nowhere with my attempts to find a viable proof to (1.).
Guess:
Let $\Lambda_n$ the set of all possible values of $X_n$. Define some $\Phi\uparrow: \mathbb{N} \to \mathbb{N}$, such that $X_{\Phi(n)}$ a subsequence of $X_n$. Chose $\Phi(n)$ such that $\forall \Phi(n)>\Phi'(n), x \in \Lambda_{\Phi(n)} \implies x\in \Lambda_{\Phi'(n)}$. One possible subsequence would be defined by $\Phi(n)=2n$.
It can be proved that for such sequences $\cup_{n \in \mathbb{N}}\Lambda_{\Phi(n)} = lim_{n \to +\infty} \Lambda_{\Phi(n)}$.
Notice that $\Lambda_n=\{\frac{1}{n}, \frac{2}{n}, ..., \frac{n}{n}\}$. Thus we know that $\forall n \in \mathbb{N}, \Sigma_{k=0}^{n}\frac{a_nk}{n^2+k^2}=1$. Re-write as $a_n\Sigma_{k=0}^{n}\frac{k}{n^2+k^2}=1.$ It's straightforward that this is also true for $\Phi(n)$.
I'm admittedly stuck here, because I think the best way would be to work further with the limits of subsequences, to show some limit-behavior of the sum $\Sigma_{k=0}^{n}\frac{k}{n^2+k^2}$ that in turn provides some information about $a_n$.
Thank you for your help - please, do not spoil (2.) in any way!
It looks to me that this is true:
$$\forall n \in \mathbb{N}, \sum_{k=1}^{n}\frac{a_nk}{n^2+k^2}=1$$