Music-inspired harmony search algorithm and its experience-based derivative

Abstract

Harmony search is a music-inspired optimization algorithm, which has been so far applied to problems in cultural fields such as Sudoku puzzle solving, music composing, and art work recognizing and to various technical fields including computer science, electrical engineering, civil engineering, mechanical engineering, and physics, obtaining solutions more efficiently while overcoming the drawbacks of existing methods. While traditional techniques depend on a calculus-based derivative or an exhaustive enumeration in order to find optimal solutions, harmony search utilizes a novel derivative based on musicians' experiences. This paper introduces various applications of harmony search and its core theory: the stochastic derivative. In addition, harmony search is compared with one of the most popular techniques, the genetic algorithm.