Researching On Multi-product Batch Production Scheduling Problem Based On Particle Swarm Algorithm

There is a large proportion of process industry in the world economy,process industry is the basic industries and pillar industries in many countries. With the changes of market demand, the production processes gradually transform to multiproduct batch process. Inherent flexibility of multiproduct batch process determines that its efficiency and effectiveness mainly depends on scheme of production planning and scheduling. However, The production scheduling problem is a complex optimization problem, it is difficulty to solve the problem by use the existing theories and methods, it is urgent to study the modeling and optimization techniques of multi-product batch scheduling. to guide the enterprise practice, reduce business operation costs, improve enterprise management level.Different production material use the same production process through the equipment in the different multi-product batch production scheduling, Similar to the classic discrete process Flow Shop scheduling problem. Based on previous study, the paper studied multi-product batch manufacture scheduling problem by using PSO algorithm. And the study would be used for aluminums production and manufacturing process to solve practical problems.Firstly, three methods were used to solve the model by transforming multi-product batch scheduling problem which was processed by a single machine in each stage consists to classic general Flow-shop scheduling problem. The former two method established mixed-integer programming model by using the traditional continuum modeling. The third method was improve Particle Swarm Optimization algorithm, It was verified that the traditional model solved small problems better, and Particle Swarm Optimization algorithm was a more efficient way for large-scale problems.Secondly, the problem was further expanded to real production environment, and multi-product batch scheduling problem coordinated by the parallel device in each stage was studied on the basis of the above study. It is more complex due to join the choice of parallel device. The model was established by transforming the problem to classic mixed Flow-shop scheduling problem. An improved PSO algorithm was proposed based on the mixed integer linear programming model established by past scholars. The ability and efficiency of solving was increased by combining with the simplex algorithm and PSO algorithm, Comparing the algorithm result and the result of classic examples from literature and optimization software, it can be concluded that the traditional model is difficult to be solved and the solution quality is not high with the increase of the size of the problem, and PSO reflects the excellent performance in tackling the large-scale practical issue.Thirdly, the modern aluminum industrial production was briefly descripted and analyzed based on above research. Production process was simplified according to actual production process. And production process from bauxite to industrial aluminum ingot in someone aluminum factory was extracted. After analysis, it was abstracted into a multiproduct batch process in which the production was coordinated by the parallel device. Then, the products of all stages were batched to calculate and determine the processing time and energy consumption of the corresponding batch. Finally, it was turned into classic hybrid Flow-shop scheduling problem. The mathematics model in which the scheduling target was minimum energy was established. Comparing and analyzing the obtained result by using improved PSO mixed with simple shaped method and basic GA results, the advantages and ability to solve practical problems of the Particle Swarm Optimization algorithm were further verified.Finally, the paper summarized the full text and looked forward to the future development and application of the studied problem.