Reinforcement learning-based alpha-list iterated greedy for production scheduling

• Reinforcement-learning-based α-list Iterated Greedy (RAIG) algorithm is developed • RAIG performs better than the benchmark considering the overall performance • Statistical results are supportive of meaningful outperformance Metaheuristics can benefit from analyzing patterns and regularities in data to perform more effective searches in the solution space. In line with the emerging trend in the optimization literature, this study introduces the Reinforcement-learning-based Alpha-List Iterated Greedy (RAIG) algorithm to contribute to the advances in machine learning-based optimization, notably for solving combinatorial problems. RAIG uses an N -List mechanism for solution initialization and its solution improvement procedure is enhanced by Reinforcement Learning and an Alpha-List mechanism for more effective searches. A classic engineering optimization problem, the Permutation Flowshop Scheduling Problem (PFSP), is considered for numerical experiments to evaluate RAIG's performance. Highly competitive solutions to the classic scheduling problem are identified, with up to 9% improvement compared to the baseline, when solving large-size instances. Experimental results also show that the RAIG algorithm performs more robustly than the baseline algorithm. Statistical tests confirm that RAIG is superior and hence can be introduced as a strong benchmark for future studies.