Application of an Iterative Greedy Algorithm with Population Strategy in Distributed Permutation Flow Shop Scheduling Problems

As the globalized world, the distributed multi-site production model has emerged as the dominant approach in the manufacturing sector, offering businesses flexible and efficient solutions for production. In actual production scenarios, processing times can vary due to a multitude of factors. Most existing studies only consider the linear deterioration process based on machine wear, whereas in reality, on one hand, the deterioration process of machines is often complex, and on the other hand, learning effect due to machine running-in and operator proficiency also need to be taken into account. Therefore, this paper studies the Distributed Permutation Flow Shop Scheduling problem considering deterioration and learning and designs a novel nonlinear deterioration and learning model that simultaneously includes both time factors and processing location factors. To enhance the solution accuracy of such a model, this paper proposes a population-based hybrid iterative greedy algorithm (PHIG). Specifically, PHIG is designed to balance population diversity and computational cost by implementing an initialization method that incorporates a pruning strategy and a balance breaking mechanism of a roulette wheel based on idle time. To enhance the algorithm's ability to explore the solution space, an adaptive perturbation operator is designed. Finally, the availability of the initialization method and the perturbation operator is verified through examples from the Taillard dataset. At the same time, the algorithm is compared with the average relative percentage deviation of three other intelligent optimization algorithms. The results indicate that this algorithm has advantages in solving proposed problem.