Packings of circular discs. How a probabilistic perspective can complete a geometric analysis (Q782893)

Let \({\mathbf x}\) be the set of centers \(x_1,\dots,x_n\) of a packing of \(n\) unit circular discs in the plane. The article deals with the problem of minimizing the quadratic energy \({\mathcal E}_n({\mathbf x})= \sum_{1\le i\le n}|x_i-\overline x|^2\), where \(\overline x\) is the centroid of \({\mathbf x}\). The first (elementary) part restricts itself to packings with centers in the usual triangular lattice. After calculating the quadratic energy for hexagonal packings, it shows that hexagonal packings minimize the quadratic energy (for a given number of discs) if and only if their side length it at most seven. This is deduced from a result of \textit{T. Y. Chow} [Combinatorica 15, No. 2, 151--158 (1995; Zbl 0832.52003)]. The second part is a short description of earlier work of the last two authors (and coauthors). Here approximate solutions of the minimization problem are obtained as equilibrium states of hard ball stochastic dynamics, described by stochastic integral equations. The balls move according to a Brownian motion, with mutual attraction and elastic collisions. The long time development of such systems supports the conjecture that the quadratic energy is minimized by packings which are circular clusters and centered in the triangular lattice.