A Novel Hybrid GWO-PS Algorithm for Standard Benchmark Optimization Problems

Abstract

Recently developed Grey Wolf Optimizer (GWO) algorithm has conspicuous behavior for verdict of global optima, without getting ensnared in premature convergence and has been applied to benchmark problems including engineering design and optimization problems. In the proposed research, the exploration phase of the Grey Wolf Optimizer has been further improved using pattern search algorithm, which is a derivative-free search method. To overcome the problem of stagnating in neighborhood optima, it involves two moves: one is pattern move and other is exploratory search. In the proposed research, a hybrid version of Grey Wolf Optimizer algorithm combined with pattern search (hGWO-PS) algorithm has been developed for the solution of various non-linear, highly constrained engineering design and engineering optimization problems. To indorse the results of the proposed hybrid algorithm, 23 benchmark problems including two real-life biomedical problems are taken into consideration. Experimentally, it has been observed that the exploitation phase in the proposed hybrid GWO-PS algorithm is better than standard Grey Wolf Optimizer algorithm, Ant Lion Optimizer algorithm, Moth Flame Optimization algorithm, sine–cosine optimization algorithm and other recently reported heuristics and meta-heuristics search algorithm. However, computational time of the algorithm has been slightly increased due to increase in the number of fitness evaluations. Hence, proposed algorithm indorses its effectiveness in the field of nature inspired meta-heuristics search algorithms.