A Bi-objective Grey Wolf Optimizer with a Dynamic Number of Leaders Applied to Spare Parts Stock Level Optimization Problems

The Grey Wolf Optimizer (GWO) is a metaheuristic optimization algorithm in the field of swarm intelligence that has been successfully used for solving a wide range of optimization problems. The GWO algorithm is based on the leadership hierarchy observed in grey wolf packs. In the original GWO, a fixed number of wolves with the best fitness values for the optimization problem under consideration is used to guide the search process. In a bi-objective optimization problem, we are interested in finding a Pareto front approximation with high cardinality, i.e. a high number of non-dominated solutions. Also, the points should be well distributed along the Pareto front. In this paper, we propose a bi-objective variant of GWO, denoted by BGWODNL. The proposed model uses a data-driven approach to define a dynamic number of leaders to guide the search process in each iteration of the algorithm. Numerical experiments are conducted considering a spare parts list optimization problem. For comparison purposes, experiments with a standard multi-objective variant of GWO are also carried out. Other benchmark multi-objective algorithms such as the Pareto Ant Colony Optimization (PACO) and the Non-dominated Sorting Genetic Algorithm (NSGA-II) are also considered. The results show that the proposed algorithm outperforms the benchmark algorithms in terms of cardinality, simulation time, and distribution of points along the Pareto front.