Opposition-based differential evolution with ordering strategy on elite individuals

A differential evolution algorithm that bases on the generating of opposition individuals, and applies individual ordering strategy on the elites is proposed. First, the opposition-based method extends the search fields to the symmetrical positions. Then all the existing individuals are sorted into two sub-populations, according to the different fitness values. Elitism ordering strategy is applied to the individuals with better fitness to improve the capability of local search, while the commonly random differential evolution method is used to the rest individuals, aiming at the diversity improvement. Simulation experiments are implemented based on a set of benchmark functions, and the result shows the promising performance of the proposed algorithm.