Production Batch Planning Problem Of Particle Swarm Optimization

The lot sizing problem attracts attention because it can bring great economic benefit and it is difficult for solving this problem in theory and algorithm. Particle swarm optimization algorithm (PSO) is swarm intelligence algorithm, which is based on the metaphor of social interaction and communication such as bird flocking. This paper makes some research which designs some algorithms based on PSO for some lot sizing problems and does some experiments in order to prove proposed algorithm's effectiveness and preponderance.The detailed contents include:The first chapter surveys the background and some related accurate, heuristic and stochastic optimization algorithms for lot sizing problem.The second chapter describes the basic principle of particle swarm optimization algorithm and its current development.The third chapter introduces cultural evolution mechanism and the structure of the cultural algorithm, brings forward parallel particle swarm optimization algorithm based on cultural evolution (PPSOCE). Select knapsack problem as experiment problem. By computing the example of other literatures and comparison of the result, it can be found that this proposed algorithm avoids particle swarm optimization algorithm's easily getting struck into premature and improves the performance of global search.The fourth chapter designs the algorithm for single level capacitated dynamic lot-sizing problem based on particle swarm optimization principle and memory mechanism. Series of test instances are solved. By computing the instance of other literatures and comparison of the result, it demonstrates the effectiveness of this algorithm.The fifth chapter designs the algorithm for group technology lot-sizing problem based on particle swarm optimization principle and memory mechanism. Series of test instances are solved. By computing the instance of other literatures and comparison of the result, it demonstrates that the algorithm of this paper is better than genetic algorithm and improved genetic algorithm.The sixth chapter first designs parallel particle swarm optimization algorithm based on cultural evolution for multi-level uncapacitated lot sizing problem, illustrates the detailed realization of the algorithm and series of test instances are solved. A parallel particle swarm optimization algorithm based on cultural evolution and a traditional particle swarm optimization algorithm are coded and used to solve the test problem in order to compare them. Experimental results show that the parallel particle swarm optimization algorithm based on cultural evolution is more effective than the traditional particle swarm optimization algorithm. This proposed algorithm is applied for multi-level capacitated lot sizing problem (MLCLSP). Constrained terms in MLCLSP are processed by the penalty function and series of test instances are solved. By computing the instance of other literatures and comparison of the result, it demonstrates the effectiveness of this algorithm.The seventh chapter designs the two-level particle swarm optimization algorithm for batch size decision-making in semiconductor wafer fabrication, which is integrated problem of lot sizing problem and scheduling problem. The first level applies PPSOCE for lot sizing problem and the second level applies traditional PSO for scheduling problem. By computing the instance of other literatures and comparison of the result, it demonstrates that the algorithm of this paper is better than ant algorithm and heuristic algorithm.The eighth chapter summarizes the work, innovation and deficiency of this paper. At the same time, it brings forward future research expectation.In combination with the Particle swarm optimization principle and cultural evolution mechanism, firstly, this paper brings forward the parallel particle swarm optimization algorithm based on cultural evolution; secondly, this paper mainly designs corresponding algorithms and solutions for some lot sizing problems. In addition, this paper also extends application area of particle swarm optimization algorithm. And from its excellent performances in various problem solutions, we can see that, with the progressing studies and improvements, the particle swarm optimization algorithm shall be applied to more and more fields.