Unconstrained polarization (Chebyshev) problems: basic properties and Riesz kernel asymptotics (Q2072147)

Let \[ K_s(x,y)=|x-y|^{-s},\;\;\;x,y\in \mathbb R^p,\;s>0, \] be the Riesz kernel. Supose that \(A\subset \mathbb R^p\). The authors consider the max-min optimization problem \[ P_s(A,\omega_N)=\inf_{y\in A}\sum_{i=1}^N K_s(x_i,y),\;\;\mathcal{P}_s^*(A,N)= \sup_{\omega_N\subset \mathbb R^p} P_s(A,\omega_N), \] where \(\omega_N\) is a list of \(N\) points in \(\mathbb R^p\) (repetition is allowed). This is an unconstrained polarization problem (in the sense that we do not demand that \(\omega_N\subset A\)). Among the main results of the article are the following: 1. Suppose that \(\omega_N^*\) is a maximal configuration and \(s>p-2\), \(p\geq 2\). The authors give an upper bound for the number of points in \(\omega_N^*\) that do not lie in an \(\varepsilon\)-neighborhood of \(A\). 2. Under regularity assumptions on \(A\), the limiting behavior of \(\mathcal{P}_s^*(A,N)\), as \(N\to \infty\), is studied in detail. Related estimates involving the size (e.g., Minkowski content) are proved. 3. Several results are proved for the important special case when \(A\) is the \((p-1)\)-dimensional sphere. \par The article contains many more results and studies more general kernels. It also contains several open problems.