The Research On Scheduling Of Batch Processing Machine Based On Hybrid Ant Colony Algorithm

With the development of information technology, computer integrated manufacturing system (CIMS) has been widely used in manufacturing enterprises. Production scheduling, as an important part of CIMS, is the main problem of Manufacturing’s production management as well as the key to the information construction. In the different production systems, the problems of production scheduling are diverse which also cause the scheduling intricately, especially in the Semiconductor. Semiconductor production system has the characteristics of processing reentrancy as well as normal processing and batch processing mixed. The processing scheduling of Semiconductor production is proved to be NP-hard problems, especially in the batch processing machines. With the complex scheduling problems, batch processing machine is becoming a hotspot for many researchers and practitioners. So research on the scheduling of batch processing machine has great theoretical and practical significance.Batch processing machine can process one or more work piece of the same product in the same process simultaneously. When it is idle, the batch processing machine need to choose work piece from the waiting queue and process immediately or wait for the new arrival workpieces. Therefore, the scheduling problems of batch processing machine can be decomposed into two sub-problem, they are the batching rule and batch sorting. In the job scheduling of batch processing machine, classic heuristic scheduling methods, in the simple and fast solving features, occupy the mainstream situation in the scheduling rules, however, due to the lack of integrity and dynamic, its accuracy is not high for solving optimization. Intelligent algorithms based on the computer technology can help to solve complex scheduling problems dynamically. Ant colony optimization algorithm (ACO) with its parallel distributed solving advantage has gradually applied to job scheduling, but it is focused on dealing with the batch ruling or batch sorting problem and lack a comprehensive study of the two sub-problem.For the above, the paper tries to make a specific study on the batch processing machine by establishing scheduling model, designing algorithm and simulation experiment. At first, this paper uses triples (X, Y, Z) to explain the particular scheduling problems of the batch processing machines, sets the total tardiness cost as the scheduling objective function, then builds a two-level structure model of the scheduling process in the batch processing machines. Secondly, concerned about the features of batch processing machines, the paper designs a hybrid ant colony algorithm based on the ant colony algorithm to deal with batching rule and batch sorting. As more and more semiconductor companies set different priority for different orders in view of the amount of orders, customer importance, and other factors of production technology. Therefore, the paper chooses multi-priority order as a prerequisite of the scheduling in batch processing machines. Finally, the paper builds two models by the simulation software, and makes multiple comparison tests between the proposed hybrid ant colony algorithm and CR, EDD, FIFO.Through a large number of experimental data collation and analysis, the results show that the proposed algorithm has a better performance in the total delay cost than heuristic rules and stable in other performance indicators. So the proposed algorithm is feasible and effective.