Parameter Tuning of A Binary Pareto Whale Optimization Algorithm Using Taguchi Grey Relational Analysis

Parameter values of any metaheuristic algorithm affect the performance of the algorithm search. However, using statistics to estimate the proper values for the algorithm’s parameters will be feasible to make optimization algorithm more robust and effective. The aim of this paper is to investigate the optimal control parameter values for the Binary Pareto Whale Optimization Algorithm BPWOA, which is used for solving maintenance scheduling problem at a power plant. Three algorithm control parameters involving population size, iteration number, and archive size were fine-tuned using Taguchi-Grey Relational Analysis GRA to achieve an optimal maintenance schedule with maximum power supply, minimum fuel expense, and minimum Carbone Dioxide CO2 emissions. The algorithm runs carried out based on Taguchi experiment design using L25 orthogonal array. The Grey Relational Grade GRG metric is utilized to evaluate the BPWOA performance. The results show that the Taguchi-Grey relational analysis approach is a dependable and efficient way to generate new optimal values for the BPWOA control parameters, allowing for multi-objective power plant maintenance scheduling with fewer runs in less time and a 20% improvement in GRG of objectives.