A Hybrid of Differential Evolution and Particle Swarm Optimization for Global Optimization

A hybrid differential evolution (HDE) approach derived from both the differential evolution (DE) and the particle swarm optimization (PSO) is proposed. In HDE, individuals in a new generation are created, not only by crossover and mutation operation as in DE, but also by PSO operations. The concepts of inertia weight and neighbor topology are adopted in HDE. The former is employed to provide consistency and diversity by adding a weighted velocity to the trial vector. In the latter, instead of the whole population, each individual can only communicate with its neighbors, and each individual creates its trial vector based on the best individual found by its neighbors so far. The proposed approach is employed for four well-known benchmarks, and the simulation results have shown its feasibility and effectiveness.