A Harmony Search Variant Based on a Novel Synthesized Approach for Constrained Numerical Optimization

• Published in: 2019 16th International Conference on Electrical Engineering, Computing Science and Automatic Control (CCE) pp. 1 - 6
• Main Authors: Barbosa-Mendez, Mario Alejandro, Portilla-Flores, Edgar Alfredo, Vega-Alvarado, Eduardo, Calva-Yanez, Maria Barbara, Sepulveda-Cervantes, Gabriel
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-09-2019
• Subjects: Benchmark testing; constraints; Harmonic analysis; Harmony Search; Linear programming; metaheuristics; numerical optimization; Optimization; Power harmonic filters; synthesizer; Synthesizers
• ISSN: 2642-3766
• DOI: 10.1109/ICEEE.2019.8884555

In this work, a novel variant of the Harmony Search algorithm called Synthesized Harmony Search is presented. It is based on the application of a synthesizer to the process of music composition simulated in the original Harmony Search. Additionally, constraints are handled by integrating the rules of Deb to the algorithm. The effect of the synthesizer over the performance is evaluated with a well-known benchmark of constrained numerical optimization problems. The proposed algorithm was compared to the original Harmony Search and one of its previously developed variants, HS_Mod2, with a statistical analysis using the best solution obtained per each algorithm in thirty independent executions. In general, the solutions generated by the Synthesized Harmony Search surpassed the corresponding values produced by the original algorithm, and were similar to those solutions from HS_Mod2, but presenting a steadier behavior.