Energy-Efficient Job Shop Scheduling Solutions: A MILP Approach

Effective job shop scheduling is crucial for optimizing productivity in tobacco manufacturing, where dynamic factors such as worker capabilities significantly impact efficiency. This study addresses the limitations of traditional scheduling models by integrating worker learning and forgetting effects into the job shop scheduling process. The primary objective is to enhance production efficiency by minimizing makespan and balancing workloads among machines and workers. We propose a mixed-integer linear programming (MILP) framework, solved using an improved memetic algorithm (IMA) that leverages adaptive neighborhood search strategies. Comparative experiments with simulated annealing (SA) and particle swarm optimization (PSO) demonstrate that the IMA achieves superior performance across multiple scenarios, particularly in minimizing makespan and energy consumption. The results show that accounting for worker learning and forgetting effects leads to more accurate predictions of production capacity and better utilization of human resources, ultimately enhancing overall productivity. This study offers a comprehensive approach to integrating human factors into industrial scheduling, providing valuable insights for both academia and industry practitioners.