An Enhanced Fruit Fly Optimization Algorithm Based on Elitist Learning and Differential Perturbation Strategy

Aim at the problem that original fruit fly optimization algorithm (FOA) has some disadvantages of low convergence precision, slow convergence rate and easily relapsing into local extremum, an enhanced fruit fly optimization algorithm based on elitist learning and differential perturbation strategy is put forward. In this study, a smell concentration based subgroup collaboration strategy is utilized, where the fruit fly swarm can be divided into excellent subgroup and general subgroup which have quite different evolutionary routes. Additionally, in order to enhance the search efficiency and keep the diversity of solutions, the elitist learning and differential perturbation strategy are adopted to coordinate the exploitation ability and the exploration ability of algorithm. Experimental results on six typical benchmark function optimization problems demonstrate the effectiveness of the algorithm as an optimization technique. Compared with original fruit fly optimization algorithm for optimization task, the presented algorithm has the ability to search better solutions and its optimization performance is clearly better than that of the original algorithm.