Iterated dynamic neighborhood search for packing equal circles on a sphere

Abstract

In this work, we investigate the equal circle packing problem on a sphere (ECPOS), which consists in packing N equal non-overlapping circles on a unit sphere such that the radius of circles is maximized. The problem is of great interest in biology, engineering and operations research and thus has a rich research history both from theoretical and computational aspects. We propose from the point of view of computational research an effective iterated dynamic neighborhood search (IDNS) algorithm for the ECPOS problem. The algorithm includes a multiple-stage local optimization method, a general dynamic neighborhood search method and an adjustment method of the minimum distance between the points on the unit sphere. Extensive experiments are conducted with the proposed algorithm on 205 instances commonly used in the literature. Computational results show that the algorithm is highly effective by improving the best-known results for 42 instances and matching the best-known results for other 116 instances, while missing the best-known results for only 5 instances. For the remaining 42 instances, the best-known results are reported for the first time by the IDNS algorithm.