A Q-learning-based biology migration algorithm for energy-saving flexible job shop scheduling with speed adjustable machines and transporters

Due to the increasing demand for green manufacturing, energy-saving scheduling problems have garnered significant attention. These problems aim to reduce energy consumption at the production system level within workshops. To simulate a realistic production environment, this study addresses an energy-saving flexible job shop scheduling problem that considers two types of speed-adjustable resources, namely machines and transporters. The optimization objective is to minimize the comprehensive energy consumption of the workshop. A novel mathematical model is initially constructed based on the specific characteristics of the problem at hand. Given its NP-hard nature, a new Q-learning-based biology migration algorithm (QBMA) is proposed, which encompasses diverse search strategies and employs a Q-learning algorithm to dynamically select search strategies, thereby preventing blind search during the evolutionary process. The experimental results of our study demonstrate the promising efficacy of QBMA in effectively addressing the aforementioned problem, while also highlighting the positive impact of considering resources with adjustable speed.