求解相同并行机混合流水线车间调度问题的分布估计算法  

Estimation of distribution algorithm for solving hybrid flow-shop scheduling problem with identical parallel machine

作　　者：王圣尧[1,2] 王凌[1,2] 许烨[1,2] 

机构地区：[1]清华大学信息科学与技术国家重点实验室,北京100084 [2]清华大学自动化系,北京100084

出　　处：《计算机集成制造系统》2013年第6期1304-1312,共9页Computer Integrated Manufacturing Systems

基　　金：国家自然科学基金资助项目(61174189;60834004);高等学校博士学科点专项科研基金资助项目(20100002110014);国家973计划资助项目(2009CB320602)~~

摘　　要：针对相同并行机混合流水车间调度问题,提出了一种有效的分布估计算法。针对基于排列的编码方式,设计了改进的启发式解码规则,进而提出了一种评价个体优劣的混合解码方式。建立了描述问题解空间分布的概率模型,通过对概率模型采样产生新个体,并基于优势种群更新概率模型的参数。通过基于标准测试集的数值仿真以及与已有算法的比较,验证了所提算法的有效性。To solve the Hybrid Flow shop Scheduling Problem with Identical Parallel Machine （HFSP-IPM , an effective Estimation of Distribution Algorithm （EDA） was proposed. Aiming at the encoded mode based on permutation, the improved heuristic decoding rules were designed, and a hybrid decoding scheme was also created. In the algorithm, a probability model for describing the space distribution of problem solution was constructed. It generated new individuals by sampling the probability model and updated the parameters of the probability model with the superior population. Simulation results based on benchmark testing problems and comparisons with some existing algorithms demonstrated the effectiveness of the proposed algorithm.

关 键 词：混合流水车间调度 相同并行机 分布估计算法 概率模型 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]