Measuring Performance of a Hybrid Optimization Algorithm on a Set of Benchmark Functions

Abstract

Hybrid algorithms are effective in solving complex optimization problems instead of using traditional methods. In literature, many proposed hybrid algorithms can be seen in order to increase their performance by the use of features of well-known algorithms. The aim of hybridization is to have better solution quality and robustness than traditional optimization algorithms by balancing the exploration and exploitation goals. This paper investigates the performance of a novel hybrid algorithm composed of Differential Evolution algorithm, Particle Swarm Optimization algorithm and Harmony Search algorithm which is called HDPH. This is done on a set of known benchmark functions. The experimental results show that HDPH has a good solution quality and high robustness on many benchmark functions. Also, in HDPH all control parameters are randomized in given intervals to avoid selecting all possible combination of control parameters in given ranges.