Differential Evolution Based Particle Swarm Optimization

A new, almost parameter-free optimization algorithm is developed in this paper as a hybrid of the barebones particle swarm optimizer (PSO) and differential evolution (DE). The DE is used to mutate, for each particle, the attractor associated with that particle, defined as a weighted average of its personal and neighborhood best positions. Results of this algorithm are compared to that of the barebones PSO, Von Neumann PSO, a DE PSO, and DE/rand/1/bin. These results show that the new algorithm provides excellent results with the added advantage that no parameter tuning is needed