文摘
The harmony search (HS) method is an emerging meta-heuristic optimization algorithm inspired by the natural musical performance process, which has been extensively applied to handle numerous optimization problems during the past decade. However, it usually lacks of an efficient local search capability, and may sometimes suffer from weak convergence. In this paper, a memetic HS method, m-HS, with local search function is proposed and studied. The local search in the m-HS is based on the principle of bee foraging like strategy, and performs only at selected harmony memory members, which can significantly improve the efficiency of the overall search procedure. Compared with the original HS method and particle swarm optimization (PSO), our m-HS has been demonstrated in numerical simulations of 16 typical benchmark functions to yield a superior optimization performance. The m-HS is further successfully employed in the optimal design of a practical wind generator.