Investigating the convergence characteristics of harmony search

Abstract

Harmony Search optimization algorithm has become popular in many fields of engineering research and practice during the last decade. This paper introduces three major rules of the algorithm: harmony memory considering (HMC) rule, random selecting (RS) rule, and pitch adjusting (PA) rule, and shows the effect of each rule on the algorithm performance. Application of example benchmark function proves that each rule has its own role in the exploration and exploitation processes of the search. Good balance between the two processes is very important, and the PA rule can be a key factor for the balance if used intelligently.