融合决策树的分布式多工厂协同生产调度方法  

Synergic Production Scheduling Method for Distributed Multi-Plants Based on Fusion Decision Tree

作　　者：王艳 蒋天伦 Wang Yan;Jiang Tianlun(Engineering Research Center of Internet of Things Technology Applications Ministry of Education,Jiangnan University,Wuxi 214122,China)

机构地区：[1]江南大学教育部物联网技术应用工程中心

出　　处：《系统仿真学报》2019年第11期2181-2197,共17页Journal of System Simulation

基　　金：国家自然科学基金(61973138)

摘　　要：在分布式多工厂协同生产调度优化问题中,需同时考虑工件在工厂间柔性分配与工件在工厂内柔性调度2个阶段的优化。建立以制造总成本与提前/延期为优化目标的分布式多工厂调度模型,提出一种融合ID3决策树的高斯粒子群优化嵌套寻优算法框架。该框架将各工厂内部独立的调度优化嵌套于工厂间分配寻优过程,并引入精英保留策略提高算法寻优性,将ID3决策树技术融入外层寻优粒子生成过程来降低外层寻优的随机性。通过仿真验证算法在寻优性、收敛性和CPU时间方面的优越性。In the synergic production scheduling optimization problem of distributed multi-plants, it is necessary to consider the two stages of job allocation between factories and job scheduling in factories at the same time. This paper first establishes a distributed multi-plant scheduling model with total cost and advance/delay as the optimization objectives, and then proposes a nested optimization algorithm framework integrating ID3 decision tree with Gauss particle swarm optimization. In this framework, independent scheduling optimization within the factory is nested in the process of inter factory allocation optimization, and elite retention strategy is introduced to improve the algorithm optimization. In addition, ID3 decision tree technology is integrated into the process of outer layer optimization particle generation to reduce the randomness of outer layer optimization. Simulation results show that the algorithm has advantages in optimization, convergence and CPU time.

关 键 词：分布式多工厂 ID3决策树 多目标粒子群优化 精英保留策略 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]