Metaheuristic Approach Hybridizing Variable Neighborhood Search and Genetic Algorithm to Size Hybrid Backup Systems for Power Outages

Abstract

This paper proposes a hybridization approach of metaheuristic techniques for sizing hybrid backup systems to supply essential loads during power outages. Two classic metaheuristic techniques were considered for this application, with contrasting search philosophies: the Genetic Algorithm and the Variable Neighborhood Search. The proposed method is analyzed through a case study for the maintenance of auxiliary services in a substation. The hybrid approach was compared with the pure metaheuristics that it is based on, observing the processing time and quality of the solution found. Finally, a statistical analysis was carried out to verify the performance of the techniques in a sequence of repetitions. The results obtained in this study suggest that although the hybridization proposal has a higher processing time than the other methods, it provided greater assertiveness in solving the problem.