Abstract: To improve the optimization performance of harmony search algorithm， an improved novel global harmony search (INGHS) algorithm was proposed. The harmony memory diversity was defined by the difference vector norm to implement dynamically position updating， and the mutation operation was employed to update harmony memory. A new harmony generated with dynamically updating strategy made the algorithm having excellent global search performance in the early states and local search capability during the final stages of the optimization process， which avoided the algorithm trapped in local optimal. Finally， the simulation for 7 benchmark functions were carried out using the proposed algorithm and other state-of-the-art HS variants that reported in the recent literature. The simulation results demonstrated that the proposed algorithm has a higher precision.

Key words: harmony search, norm, diversity, adaptive, mutation