 |
In this paper, we address the integrated process planning and scheduling problem (IPPS) with reconfigurable machine tools (RMTs). For solving the problem, we adopt an online approach where both decisions are integrated, thus simultaneously taken. We propose a novel metaheuristic, Q-NSGA-II, which combines the Non-Dominated Sorting Genetic Algorithm (NSGA-II) with Q-learning. This approach dynamically adjusts the crossover probability after each iteration using Q-learning, aiming to improve the search process. The proposed metaheuristic was compared to three standard NSGA-II variants with fixed crossover probabilities (0.1, 0.5, and 0.9). The four metaheuristics were tested for solving an instance from the literature, and the results demonstrated that Q-NSGA-II provides better results than its static counterparts in terms of quality and stability. This suggests that Q-NSGA-II has the potential to be a strong competitor in the IPPS domain, warranting further investigation and comparison with other existing solution approaches.
