关键词: 智能车间, 设施布局, k-平行行排序问题, 物料搬运通道, Q-learning

Abstract: Aiming at the deficiency of the coordinated optimization between multi-row layouts and vertical material handling passages,a k-parallel row ordering problem with multi-handling passages for smart workshops was proposed.A mixed-integer linear programming model with a logistics cost minimization objective was established,where the Gurobi solver was employed for small-scale instances.For large-scale problems,a Q-learning-enhanced variable neighborhood search genetic algorithm was proposed.The algorithm enhanced initial population quality using a greedy strategy by integrating with the problem features,and the state space was designed,and designs the action space with multiple neighborhood search operations was introduced.An ε-greedy policy ensures adaptive exploration.Benchmark tests on 9~60 scale instances demonstrated the algorithm's superiority.Finally,practical workshop case validation confirmed the method's effectiveness in intelligent workshop layout optimization,providing theoretical and practical support for digital-intelligent transformation in manufacturing workshops.

Key words: smart workshop, facility layout, k-parallel row ordering problem, material handling passages, Q-learning