基于Petri网与遗传算法的可重入生产系统调度

可重入生产系统调度问题属于NP难题,该文建立了系统的扩展Petri网模型,并且将遗传算法方法与调度规则结合起来用于解决可重入生产系统的调度问题。针对可重入生产系统生产过程的动态复杂性,首先建立了一类综合调度规则,然后提出了基于设备分组与分时段的综合规则组合的可重入生产系统调度策略,并采用遗传算法与基于Petri网模型的过程仿真相结合的方法对综合规则组合进行优化,仿真比较验证了该调度策略的有效性。     

基于综合规则与遗传算法的可重入生产系统调度

可重入生产系统调度问题属于NP难题,该文将遗传算法方法与调度规则结合起来用于解决可重入生产系统的调度问题。首先针对系统的可重入性建立了一类综合规则,然后在设备分组与分时段的基础上,提出了基于一类综合规则组合的可重入生产系统调度策略,并采用遗传算法与过程仿真相结合的办法对综合规则组合进行优化,仿真比较验证了该调度策略的有效性。     

可重入混合流水车间调度的拉格朗日松弛算法

为了有效提升多重入车间的生产效率,考虑了实际生产中检查和修复过程对于逐层制造的可重入生产系统的重要性,提出了基于拉格朗日松弛算法的可重入混合流水车间的调度方法.首先进行了问题域的描述,并在此基础上以最小化加权完成时间为调度目标,建立数学规划模型.针对该调度问题提出了基于松弛机器能力约束的拉格朗日松弛算法,使松弛问题分解成工件级子问题,并使用动态规划方法建立递归公式,求解工件级子问题.随后,使用次梯度算法求解拉格朗日对偶问题.最后,对各种不同问题规模进行了仿真实验,结果表明,所提出的调度算法能够在合理的时间内获得满意的近优解.     

可重入生产系统的平均报酬型强化学习调度

在可重入生产系统中,一个重要的问题就是对调度策略进行优化,以提高系统平均输出率.本文采用了一种平均报酬型强化学习算法来解决该问题,直接从所关心的系统品质出发,自动获得具有自适应性 的动态调度策略.仿真结果表明,其性能优于两种熟知的优先权调度策略.     

基于遗传算法的可重入半导体生产线的调度

半导体生产线的可重入性使得半导体生产车间的调度成为NP难题.基于遗传算法的调度是将遗传算法应用到半导体生产线的投料策略和调度策略上,通过合理的编码来寻找一组投料策略和调度策略的组合,来解决可重入半导体生产线的调度问题.和以往的方法不同,这种方法可以从全局上来调度半导体生产线,充分发挥出各种策略的优势.仿真结果表明这种方法对于产出量、制造周期偏差、瓶颈设备利用率等关键生产指标都有明显改善.     

可重入生产系统的递阶增强型学习调度

对平均报酬型马氏决策过程,本文研究了一种递阶增强型学习算法;并将算法应用 于一个两台机器组成的闭环可重入生产系统,计算机仿真结果表明,调度结果优于熟知的两 种启发式调度策略．     

可重入式生产车间调度的计算机仿真与优化研究

对于冷拔钢管生产车间的可重入生产的调度问题,采用计算机仿真软件系统eM-Plant来进行仿真优化求解.首先设定工件的成组规则,简化调度问题为投料策略与派工策略,然后采用遗传算法进行调度参数优化.仿真结果的对比表明遗传算法对此类问题的求解有较强的优化性能和稳定性.     

用神经网络求解机器人动态调度中的近似指派问题

本文利用HOPFIELD神经网络，对机器人动态调度中的近似指派问题提出了合理的神经网络表示方法，给出了网络的能量函数表示法及神经元状态方程，从而得出了机器人动态调度中近似指派问题的快速求解策略，满足了动态调度的实时性要求．本文从理论上论证了所提算法的收敛性．软件仿真结果表明，本文提出的近似指派问题网络求解方法是有效的，计算结果是满意的．     

离散事件和混杂系统的调度控制*

从系统控制的角度研究了离散事件和混杂系统的调度问题.回顾了我们以前的工作,包括拉格朗日松弛技术,分布实时set-up 调度,随机和可维修系统的调度,神经网络及其它搜索技术的调度,制造系统和库存系统的安全点策略,以及并行处理系统的一些新的启发调度规则,等等.然后介绍了可重入生产系统调度控制的一些新结果.可重入生产系统在现代VLSI制造系统中是非常重要的.建立了QBD型模型,并用Neuts的矩阵几何方法求解.用随机搜索方法解决了监测站的设置问题.数值仿真结果证明了这种方法的有效性和适应性.     

基于总加权完成时间的可重入混合流水车间调度问题

研究以最小化总加权完成时间为目标的可重入混合流水车间调度问题(RHFS-TWC),并构建问题的整数规划模型.根据模型的特点,设计基于二维矩阵组的调度解编码方案,结合NEH启发式算法确定工件初始加工顺序,生成高质量初始调度解群.为避免算法陷入早熟及扩大解的搜索空间,给出IGA的遗传参数自适应调整策略,最终形成NEH-IGA融合求解策略.针对不同规模问题分别用传统GA、基于遗传参数自适应调整的IGA、NEH启发式、NEH-IGA算法进行仿真测试,仿真结果表明NEH启发式和遗传参数自适应动态调整策略的引入有效改善了原有GA的求解能力,NEH-IGA算法在求解RHFS-TWC问题方面优势明显.     

Simulation Framework of Scheduling Re-entrant Lines with Nuero-Dynamic Programming

I. INTRODUCTION With the rapid development of semiconductor manufacturing, its particular re-entrant feature brings complicated scheduling problems, which have become the focus of research. After semiconductor manufacturing is abstracted as re-entrant lines by Kumar[1], various scheduling schemes are presented by scholars. Because of the shortages of scheduling methods of heuristic algorithm, dynamic scheduling methods have attracted much attention. The classical dynamic scheduling metho…     

Goal programming approach to solving network design problem with multiple objectives and demand uncertainty

The transportation network design problem (NDP) with multiple objectives and demand uncertainty was originally formulated as a spectrum of stochastic multi-objective programming models in a bi-level programming framework. Solving these stochastic multi-objective NDP (SMONDP) models directly requires generating a family of optimal solutions known as the Pareto-optimal set. For practical implementation, only a good solution that meets the goals of different stakeholders is required. In view of this, we adopt a goal programming (GP) approach to solve the SMONDP models. The GP approach explicitly considers the user-defined goals and priority structure among the multiple objectives in the NDP decision process. Considering different modeling purposes, we provide three stochastic GP models with different philosophies to model planners’ NDP decision under demand uncertainty, i.e., the expected value GP model, chance-constrained GP model, and dependent-chance GP model. Meanwhile, a unified simulation-based genetic algorithm (SGA) solution procedure is developed to solve all three stochastic GP models. Numerical examples are also presented to illustrate the practicability of the GP approach in solving the SMONDP models as well as the robustness of the SGA solution procedure.     

Genetic algorithm using sequence rule chain for multi-objective optimization in re-entrant micro-electronic production line

Re-entrant production lines, such as those which occur in micro-electronic wafer fabrication, are characterized by a product routing that consists of multiple visits to a facility during the manufacturing process. With the development of micro-electronic technology, the research on the scheduling and control problem of re-entrant micro-electronic production line has attracted more and more people from both academia and industry to study and has become a challenging research subject. Some results of the scheduling of re-entrant micro-electronic production line based on heuristic sequence rules have been obtained. However, performances of these sequence rules are not good enough in re-entrant micro-electronic production line because of their sensitivity to the variation of types of production line. A genetic algorithm using sequence rule chain for multi-objective optimization in re-entrant micro-electronic production line is proposed in this paper. Comparisons between the proposed algorithm and some other typical sequence rules have been made through the simulations of a practical micro-electronic production line. The static and dynamic simulation results show that the algorithm has considerable improvements on performances of the micro-electronic production such as mean cycle time, mean number of work-in-process, production rate.     

基于遗传算法的某航空发动机装配车间优化调度

航空发动机装配车间装配生产线的调度问题,是一类比较典型的混合Flowshop问题,同时还带有工件可重人等特点,这就区别于一般的Flowshop和Jobshop调度问题,因此,将可重入混合车间调度问题划为第三类调度问题。关于重入式混合车间生产调度的优化问题通常来说都是属于NP难问题。文中通过某航空发动机装配车间生产线的研究,以最小化最大完工时间为目标函数,借助随机矩阵的编码方式和改进的交叉方法与变异方法,提出了基于遗传算法的调度优化方法。最后实验结果表明,文中提出的改进算法能够有效地实现装配车间调度的优化。     

A sufficient condition for instability of buffer priority policies in re-entrant lines

We use a buffer-boundedness approach to study the stability of re-entrant lines with a buffer priority scheduling policy. Using Petri net models we establish a sufficient condition for instability of such systems having a positive feedback loop. An example of unstable systems is also given.     

Fuzzy neural network-based rescheduling decision mechanism for semiconductor manufacturing

Most semiconductor manufacturing systems (SMS) operate in a highly dynamic and unpredictable environment. The production rescheduling strategy addresses uncertainty and improves SMS performance. The rescheduling framework of SMS is presented as layered scheduling strategies with an optimization rescheduling decision mechanism. A fuzzy neural network (FNN) based rescheduling decision model is implemented which can rapidly choose an optimized rescheduling strategy to schedule the semiconductor wafer fabrication lines according to current system disturbances. The mapping between the input of FNN, such as disturbances, system state parameters, and the output of FNN, optimal rescheduling strategies, is constructed. An example of a semiconductor fabrication line in Shanghai is given. The experimental results demonstrate the effectiveness of proposed FNN-based rescheduling decision mechanism approach over the alternatives such as back-propagation neural network (BPNN) and multivariate regression (MR).     

Fluctuation smoothing policies are stable for stochastic re-entrant lines

In this paper we establish that all the policies in the class of Fluctuation Smoothing Policies for Mean Cycle Time (FSMCT), proposed in (Lu et al. 1992) and (Lu and Kumar 1993), are stable, in the sense of positive Harris recurrence, for all stochastic re-entrant lines. These constitute the first analytical results on this promising class of scheduling policies which have been shown to substantially reduce the mean and variance of cycle times in stochastic re-entrant lines. Stability is established by showing that the fluid limits corresponding to the FSMCT policies empty in finite time from all suitable initial conditions. A special case of FSMCT policies is the Last Buffer First Serve policy, and its stability is also thus established.     

Design of a decision support system for detailed scheduling

Design of a decision support system ( ) based on a simulation model of the detailed scheduling activities in a tractor manufacturing company is dealt with. The system analysis phase of the design process is overviewed briefly. The main decision points involved and the problems faced in the production planning and control subsystem are presented. Expectations from a for detailed scheduling are discussed and performance measures are defined. The links between computer programs are shown. Utilization of the designed for production planning and control oriented decision making is discussed using decision tables.