A Multi-point Stochastic Search Method for Global Optimization

摘要

A new multi-point stochastic search method is proposed for global optimization. The underlying idea is to mimic the physics of electromagnetism by considering each sample point as an electrical charge. Optimal (near-optimal) solutions are found through specially designed attraction and repulsion mechanisms. The potential of the proposed method is demonstrated with some test results. It is shown that in the limit a version of the proposed method converges to the vicinity of global optimum with probability one.