启发式Hopfield神经网络在车间调度中的应用  

Heuristic Algorithm Combineotion with Hopfield Neural Network Application of Job-Shop Scheduling

作　　者：黄明[1] 宫鹏[1] 梁旭[1] 

机构地区：[1]大连铁道学院电气信息分院,辽宁大连116028

出　　处：《大连铁道学院学报》2004年第1期44-47,共4页Journal of Dalian Railway Institute

基　　金：辽宁省自然科学基金资助项目(20022114)

摘　　要：调度问题是一类非常复杂的组合优化问题,而Hopfield神经网络通常被广泛应用于各种组合优化问题 针对车间调度问题(JSP)的约束条件和换位矩阵,提出了包含所有约束条件的计算能量函数表达式,并针对神经网络依赖初始解,提出了启发式算法与神经网络相结合的方法,并得到解决车间调度的Hopfield神经网络结构和权值解析表达式Job shop scheduling is a complicated combined optimization problem which Hopfield neural network is extensively applied. All constraints of Job shop scheduling problems and its permutation matrix expression are proposed. Because NN depends on the initial values, a new combinaed heuristic algorithm with NN is given. Besides, computational energy function including all constraints of job shop scheduling problem is also presented. Accordingly, new neural network construction and its weights of job shop scheduling problem are obtained. Simulation results demonstrate that the derived energy function works effectively.

关 键 词：启发式算法 HOPFIELD神经网络 车间调度 换位矩阵 能量函数 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]