Hyper-heuristic Approach With Clustering For Inter-cell Scheduling

As market demands become increasingly diverse,the traditional cell production mode in manufacturing cannot fully meet the practical production needs of "multi-variety and variable-batch".Therefore,allowing Inter-cell collaboration has become a solution that meets practical needs.The resulting Inter-cell scheduling problem refers to the coordination and scheduling of resources across multiple cell,which typically has the characteristics of large scale and high complexity.Hyper-heuristic algorithms,which manage or manipulate a series of low-level heuristic algorithms through high-level heuristic methods,have strong adaptability and high computing power,and have shown excellent performance in solving large-scale complex optimization problems.This thesis focuses primarily on using hyper-heuristic algorithms to solve Inter-cell scheduling problems,and incorporates clustering algorithms to enhance the optimization performance of the algorithm.The specific research results are as follows:(1)The dissertation proposes a hyper-heuristic algorithm based on the K-means clustering approach to solve the Inter-cell scheduling problem.Firstly,the relevant attributes of entities are determined,and the scheduling problem model is established.Then,clustering is performed based on the entity attributes to construct decision blocks,which serve as the basic working cell of the hyper-heuristic algorithm.Finally,ant colony algorithm is employed to search for heuristic rules for the decision blocks,and the rules are applied to generate scheduling solutions.Simulation experiments show that the addition of clustering ideas effectively improves the optimization performance of the algorithm.(2)A hyper-heuristic algorithm based on clustering multiple attributes of entities is proposed.Firstly,select multiple attributes related to the job entities based on the Inter-cell scheduling problem and establish a multi-dimensional data model to increase the differentiation between entities in different decision blocks.Then,using the clustering effect as the evaluation criterion,dynamically select the silhouette coefficient to optimize the number of decision blocks,enabling the algorithm to search the entire solution space within a reasonable time.Simulation results show that the hyper-heuristic algorithm based on multi-attribute clustering has better optimization performance.In summary,this dissertation addresses the Inter-cell scheduling problem,extracts the problem model based on actual production situations,and combines the hyper-heuristic algorithm with the clustering algorithm.The ant colony algorithm is used as the high-level search algorithm,and the heuristic rule is used as the low-level heuristic search strategy.During the search process,the clustering idea is applied to dynamically construct decision blocks,effectively improving the optimization performance and computational efficiency of the algorithm,and further expanding the application scenarios of clustering algorithms.