Performance evaluation of an improved harmony search algorithm for numerical optimization: Melody Search (MS)

Melody Search (MS) Algorithm as an innovative improved version of Harmony Search optimization method, with a novel Alternative Improvisation Procedure (AIP) is presented in this paper. MS algorithm mimics performance processes of the group improvisation for finding the best succession of pitches within a melody. Utilizing different player memories and their interactive process, enhances the algorithm efficiency compared to the basic HS, while the possible range of variables can be varied going through the algorithm iterations. Moreover, applying the new improvisation scheme (AIP) makes algorithm more capable in optimizing shifted and rotated unimodal and multimodal problems than the basic MS.In order to demonstrate the performance of the proposed algorithm, it is successfully applied to various benchmark optimization problems. Numerical results reveal that the proposed algorithm is capable of finding better solutions when compared with well-known HS, IHS, GHS, SGHS, NGHS and basic MS algorithms. The strength of the new meta-heuristic algorithm is that the superiority of the algorithm over other compared methods increases when the dimensionality of the problem or the entire feasible range of the solution space increases.

► Melody Search, an innovative improved version of Harmony Search, is described. ► Utilizing several memories increases the stochastic search efficiency. ► Novel improvisation procedure is introduced and adopted in this study. ► Algorithm performance is considered using eighteen numerical benchmarks. ► Obtained results are compared whit those of six other HS based methods.