Abstract: To improve the optimization performance of harmony search algorithm， a multiple-sub-groups hybrid harmony search (MHHS) algorithm is proposed. The algorithm sorts the entire harmony individuals according to the space distance between each harmony and the best harmony， and then builds multiple layers based on the ranked results， where each layer is as a unique sub-group. Different sub-groups integrate various differential adjustment strategies to broaden search ranges. Meanwhile， the communication mechanism is built to facilitate the information exchange among the multiple-sub-groups and to promote multiple-sub-groups coevolution. Simulation results show that the proposed algorithm is better than the existing algorithms such as HS， EHS， NGHS， MPSO， CLPSO， DE， ODE and IABC in terms of optimization accuracy， convergence and robustness.

Key words: harmony search algorithm, sub-group, differential adjustment, communication mechanism