求解作业车间调度问题的差分和声搜索算法  

Differential evolution harmony search algorithm for solving job-shop scheduling problem

作　　者：张敬敏[1] 李霞[1] 

机构地区：[1]石家庄经济学院信息工程学院,石家庄050031

出　　处：《计算机应用》2013年第2期329-332,356,共5页journal of Computer Applications

基　　金：河北省科技攻关项目(11213525D)

摘　　要：为能够应用和声搜索算法(HSA)高效求解作业车间调度问题(JSSP),提出一种新的差分和声搜索算法(DEHSA)。首先,针对和声函数连续而工序离散现象,设计了排序工序数量转换法,将浮点数的和声转换成工件序列;其次,为提高HSA的收敛速度,改进了HSA的进化模式,不仅是替换一个最差解,还提出了和声变量进化时依赖于当前最优解的"导优"概率;最后,将差分进化算法(DEA)引入到HSA中,克服了HSA方向性差和后期停滞的现象。仿真实验结果表明,DEHSA在求解JSSP上具有可行性和有效性。To solve the Job-Shop Scheduling Problem （JSSP） efficiently, a Differential Evolution Harmony Search Algorithm （DEHSA） was put forward. First of all, the sorting process number conversion method was designed for converting floating-point numbers harmony into a workpiece sequence to solve the phenomenon that the harmony function is continuous while the process is discrete. Secondly, in order to improve the convergence rate of HSA, its evolution model was improved instead of replacing a worst solution only, and the probability of the harmonies variable evolution depending on current optimal solution named ＂guide excellent＂ was also proposed. At last, the Differential Evolution Algorithm （DEA） was introduced to HSA to overcome the poor directional and late stagnation. A large number of simulation results show that DEHSA has good feasibility and effectiveness in job-shop scheduling problem.

关 键 词：组合优化问题 作业车间调度 和声搜索算法 差分进化算法 排序 概率 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]