A novel Particle Swarm Optimization with special boundaries restriction strategy

摘要

In this paper, the performance of the recently proposed unified Particle Swarm Optimization(PSO) with a novel boundaries restriction (BR) strategy is investigated on unstrained optimization problems. The BR strategy is to preserve feasibility of the encountered solutions in the process of different PSOs. The algorithms are illustrated on four well-known problems with promising results. Comparisons with the standard local and global variant of the methods are reported and discussed. To exhibit the performance of the new technique, simulations on the famous question min V5(X,α) for distributing 5 points on the sphere are done. Numerical results are obtained, and the minimum energy of Thomsons problem min V5(X, 1) = 6.47469149468816 is obtained.