A Chaotic Grey Wolf Optimizer Applied to Condition-Based Maintenance Optimization

Abstract

The Grey Wolf Optimizer (GWO) algorithm is a nature-inspired population-based metaheuristic that simulates the social behavior observed in a grey wolf pack. GWO has been successfully applied to different optimization problems. In this paper, we propose a chaotic version of GWO, denoted by CGWO, that uses a chaotic variable to define the number of wolves in the pack that will act as leaders, i.e. the number of wolves that guide the hunting process in each iteration of the algorithm. The proposed algorithm is used to find the optimal maintenance scope for a series-parallel system. We assume that a Prognostics and Health Management (PHM) system is available and provides the degradation level and the Remaining Useful Life (RUL) prediction for each component. The goal is to find the maintenance scope that minimizes the expected total cost per cycle until the next maintenance intervention. The performance of the proposed model is compared with the performance of the original GWO and the well-studied Ant Colony Optimization algorithm (ACO). Different chaotic maps were tested. The results show that the proposed model presented a competitive performance.