An immune-based multi-agent system for flexible job shop scheduling problem in dynamic and multi-objective environments

In this paper, a biological immune-based multi agent system (IMAS) is developed to overcome the multi-objectivity and dynamicity of the multi-objective dynamic flexible job shop scheduling (MODFJSS) problem. In modeling the MODFJSS problem, four objectives, i.e., makespan, total weighted tardiness, maximal machine workload, and schedule stability, are considered for simultaneous optimization. In the proposed IMAS architecture, immune agents are coordinated to identify the environment, generate a set of non-dominated schedules, and select a utility optimal schedule. In addition, these agents have the ability of self-adaptation and flexible coordination to adapt to the environment of MODFJSS problem. The performance of the proposed method was evaluated on dynamic versions of the five commonly used benchmark instance sets in comparison with three state-of-the-art algorithms. The obtained results showed that the proposed IMAS outperformed the competitors in most of the experiments conducted in this study.