Cosine Harmony Search (CHS) for Static Optimization

Abstract

Harmony Search (HS) is a behaviour imitation of a musician looking for the balance harmony. HS suffers to find the best parameter tuning especially for Pitch Adjustment Rate (PAR). PAR plays a crucial role in selecting historical solution and adjusting it using Bandwidth (BW) value. However, PAR in HS requires to be initialized with a constant value at the beginning step. On top of that, it also causes delay in convergence speed due to disproportion of global and local search capabilities. Even though, some HS variants claimed to overcome that shortcoming by introducing the self-modification of pitch adjustment rate, some of their justification were imprecise and required deeper and extensive experiments. Local Opposition-Based Learning Self-Adaptation Global Harmony Search (LHS) implements a heuristic factor, η for self-modification of PAR. It (η) manages the probability for selecting the adaptive step either as global or worst. If the value of η is large, the opportunity to select the global adaptive step is high, so the algorithm will further exploit for better harmony value. Otherwise, if η is small, the worst adaptive step is prone to be selected, therefore the algorithm will close to the global best solution. In this paper, regarding to the HS problem, we introduce a Cosine Harmony Search (CHS) by incorporating embedment of cosine and additional strategy rule with self-modification of pitch tuning to enlarge the exploitation capability of solution space. The additional strategy employs the η inspired by LHS and contains the cosine parameter. We test our proposed CHS on twelve standard static benchmark functions and compare it with basic HS and five state-of-the-art HS variants. Our proposed method and these state-of-the-art algorithms executed using 30 and 50 dimensions. The numerical results demonstrated that the CHS has outperformed with other state-of-the-art in accuracy and convergence speed evaluations.