4-Rule Harmony Search Algorithm for Solving Computationally Expensive Optimization Test Problems

Abstract

This paper proposes an enhanced harmony search algorithm for solving computationally expensive benchmarks widely used in the literature. We explored the potential and applicability of the original harmony search (HS) algorithm through introducing an extended version of the algorithm integrated with a new dynamic search equation enabling the algorithm to take guided larger steps at the beginning of the search. In the 4-Rule Harmony Search (4RHS), an extra rule is added to the standard HS without adding any user parameters to existing initial parameters. The 4RHS algorithm is then tested through optimal solving of different well-known and well-used benchmarks from classical to so CEC’ 2015 series, where the results of the 4RHS are compared with simple and improved version of HS algorithms as well as other optimization techniques. The obtained optimization results show the attractiveness of the added rule into the standard HS.