The setup is: consider a shape $S\in R^d$, and let $\mu_S$ be uniform measure on $S$. I want to find (among all shape with volume 1) the shape that minimizes $\mathbb{E}_{X,Y \sim \mu_S i.i.d}[\|X-Y\|^k]$.
I came across this problem in one step of my final project for a class. Intuition tells me it would be a ball since the ball is a convex shape and doesn't have sharp corners, but I just cannot come up with a rigorous proof. It has a isoperimetric inequality flavor... I think someone must have looked into this problem before, I wonder can anyone point me to some potential resources? Thanks a lot!