基于强化学习的模式驱动调度系统研究

目前,还没有一种调度规则能够根据系统环境状态的改变来进行自适应调整.对此,提出一种基于智能体的模式驱动调度系统,由智能体和仿真环境两个主要部分构成.其中,智能体将利用强化学习(Q学习算法)进行训练,以提高其动态选择合适调度规则的能力.仿真结果表明,这种模式驱动调度系统能够很好地根据系统环境状态的改变选择出对应的最优调度规则,且其调度性能优于单一调度规则,适合于系统环境不断变化的动态调度.     

Non-identical parallel machine scheduling with fuzzy processing times using genetic algorithm and simulation

This paper addresses non-identical parallel machine scheduling problem with fuzzy processing times (FPMSP). A genetic algorithm (GA) approach embedded in a simulation model to minimize maximum completion time (makespan) is proposed. The results are compared with those obtained by using longest processing time rule, known as the most appropriate dispatching rule for such problems. This application illustrates the need for efficient and effective heuristics to solve FPMSPs. The proposed GA approach yields good results and reaches them fast and several times in one run. Moreover, due to its advantage of being a search algorithm, it can explore alternative schedules providing the same results.     

飞机复合材料部件装配调度算法研究

以优化飞机复合材料部件装配顺序、缩短部件装配时间为目标,构建了复合材料部件装配调度模型,提出了一种启发式算法与禁忌搜索算法相结合的调度算法(HTA-AJSP),并对其进行了优化,通过采用变邻域搜索和禁忌规则,很好地避免了算法搜索陷入局部最优。通过分析与实例验证,并与启发式算法和局部搜索算法进行比较,所提出的算法在产品装配调度优化问题上取得了比较满意的效果,缩短了装配周期,提高了装配效率。     

飞机复合材料部件装配调度算法研究

以优化飞机复合材料部件装配顺序、缩短部件装配时间为目标,构建了复合材料部件装配调度模型,提出了一种启发式算法与禁忌搜索算法相结合的调度算法（HTA-AJSP）,并对其进行了优化,通过采用变邻域搜索和禁忌规则,很好地避免了算法搜索陷入局部最优。通过分析与实例验证,并与启发式算法和局部搜索算法进行比较,所提出的算法在产品装配调度优化问题上取得了比较满意的效果,缩短了装配周期,提高了装配效率。     

求解作业车间调度问题的快速启发式算法

首先将作业车间调度问题转换为一个搭积木模型,受这个直观模型的启发,提出了一个启发式的搭积木规则,该规则综合考虑了已经搭好的积木的顶高和将要搭积木的剩余高度。基于这个规则,提出了一个求解作业车间调度问题的快速启发式算法,对国际上通用的benchmark例的模拟实验结果表明,提出的算法优于经典的优先分配启发式算法。     

一种新型的作业车间调度算法的研究与实现

将工件的剩余加工时间分为相对剩余加工时间和绝对剩余加工时间 ,提出了优先分配启发式算法的一种新的优先分配规则 ,即相对剩余加工时间最大的概念 ,把调度分成多个阶段的部分调度 ,通过比较部分调度集合中的可调度工序的相对剩余加工时间 ,求解出每个部分调度的最优解 ,从而使整个调度达到全局最优或近似最优。最后 ,开发出调度软件 ,验证了算法在工程中的可行性、有效性。     

求解液压阀块加工车间调度的多作用力微粒群算法

为有效地解决液压阀块加工车间调度问题,考虑工序间和机器间的约束关系,以最大完成时间最小为目标,给出了液压阀块加工车间调度优化模型。为平衡算法的全局和局部搜索能力,提出了多作用力微粒群(MFPSO)算法,采用多作用力阶段性搜索策略,将搜索过程划分为前期、中期、后期3个阶段,并对应构造单一斥力、平衡引斥力、单一引力3种作用力规则,在不同搜索阶段采用不同的作用力规则,提高了算法的搜索机制和寻优性能。将MFPSO算法用于求解液压阀块加工车间调度问题,利用矩阵变量来处理约束条件,给出了一种基于矩阵的微粒编码、解码方法。通过液压阀块加工车间调度优化实例,将MFPSO算法与微粒群算法、中值导向微粒群算法、扩展微粒群算法、蚁群算法进行了对比,结果表明,提出的MFPSO算法结果最优,从而验证了该算法的有效性。     

A fuzzy dispatching strategy for due-date scheduling of FMSs with information delays

Conventional dispatching strategies for FMSs with routing flexibility have typically employed simple heuristics such as work-in-next-queue (WINQ) and number-in-next-queue (NINQ). The effectiveness of these heuristics, however, deteriorates in FMSs whose operational environment must cope with information delays that are non-negligible in comparison to part processing times. Such delays could arise from planned activities, e.g., acquisition, selection, processing, and transfer of plant-wide system status information as well as from unplanned events such as ERP/IT system malfunctions, mismatch of software interfaces, and erroneous inventory master files, for example. Uncertainties from information delays make a strong case for the introduction of fuzzy controllers for making scheduling decisions. This paper introduces a novel fuzzy logic-based dispatching strategy to cope with a specific manifestation of information delays, called status review delay within FMSs. Status review information delays impact system performance adversely because of the obsolescent nature of the information used in the determination of dispatch decisions. A fuzzy dispatching strategy (FDS), designed specifically for deployment within FMSs where information delays are manifest, provides an appropriate alternative to conventional dispatching strategies such as WINQ and NINQ. In the design of an FDS, relevant system-based parameters are fuzzified and an appropriate rule base is designed. Simulation experiments demonstrate the superiority of an FDS over the conventional WINQ dispatching strategy using the mean tardiness, percent tardy, and mean flowtime performance measures.     

A novel hybrid genetic algorithm for the open shop scheduling problem

In this paper, a hybrid genetic algorithm is proposed for the open shop scheduling problem with the objective of minimizing the makespan. In the proposed algorithm, a specialized crossover operator is used that preserves the relative order of jobs on machines and a strategy is applied to prevent from searching redundant solutions in the mutation operator. Moreover, an iterative optimization heuristic is employed which uses the concept of randomized active schedules, a dispatching index based on the longest remaining processing time rule and a lower bound to further decrease the search space. Computational results show that the proposed algorithm outperforms other genetic algorithms and is very competitive with well-known metaheuristics available in the literature.     

Information-Based Dynamic Manufacturing System Scheduling

Information about the state of the system is of paramount importance in determining the dynamics underlying manufacturing systems. In this paper, we present an adaptive scheduling policy for dynamic manufacturing system scheduling using information obtained from snapshots of the system at various points in time. Specifically, the framework presented allows for information-based dynamic scheduling where information collected about the system is used to (1) adjust appropriate parameters in the system and (2) search or optimize using genetic algorithms. The main feature of this policy is that it tailors the dispatching rule to be used at a given point in time to the prevailing state of the system. Experimental studies indicate the superiority of the suggested approach over the alternative approach involving the repeated application of a single dispatching rule for randomly generated test problems as well as a real system. In pa ticular, its relative performance improves further when there are frequent disruptions and when disruptions are caused by the introduction of tight due date jobs and machine breakdown—two of the most common sources of disruption in most manufacturing systems. From an operational perspective, the most important characteristics of the pattern-directed scheduling approach are its ability to incorporate the idiosyncratic characteristics of the given system into the dispatching rule selection process and its ability to refine itself incrementally on a continual basis by taking new system parameters into account.     

Bottleneck-based heuristic dispatching rule for optimizing mixed TDD/IDD performance in various factories

Most research on scheduling problems focuses on increasing production efficiency. For instance, the shortest processing time (SPT) and earliest due date (EDD) dispatching rules perform well in minimizing mean flow time and reducing maximum tardiness, respectively. However, those indices ignore the financial impact (material cost and order price) on the factory. Previous studies focused mainly on cycle time and due date. However, the theory of constraint (TOC) considers not only the effect of time, but also financial factors. Therefore, TOC addresses the concepts of throughput-dollar-day (TDD) and inventory-dollar-day (IDD). The former index (TDD) represents penalties for tardy deliveries, while the latter index (IDD) refers to the material holding cost. Based on these two indices, this investigation creates a novel mixed TDD/IDD weighted value (Z _value) to replace the other traditional indices for taking measurements in various factories. This study also designs a heuristic dynamic scheduling algorithm (mixed TDD/IDD dispatching rule) for reducing the system Z _value. Some traditional dispatching rules are compared with the proposed rule in terms of TDD, IDD, and Z _value. Analytical results indicate that the mixed TDD/IDD dispatching rule is feasible and generally outperforms other conventional dispatching rules in terms of Z _value under various factories.     

A dispatching rule-based hybrid genetic algorithm focusing on non-delay schedules for the job shop scheduling problem

Job shop scheduling is an important decision process in contemporary manufacturing systems. In this paper, we aim at the job shop scheduling problem in which the total weighted tardiness must be minimized. This objective function is relevant for the make-to-order production mode with an emphasis on customer satisfaction. In order to save the computational time, we focus on the set of non-delay schedules and use a genetic algorithm to optimize the set of dispatching rules used for schedule construction. Another advantage of this strategy is that it can be readily applied in a dynamic scheduling environment which must be investigated with simulation. Considering that the rules selected for scheduling previous operations have a direct impact on the optimal rules for scheduling subsequent operations, Bayesian networks are utilized to model the distribution of high-quality solutions in the population and to produce the new generation of individuals. In addition, some selected individuals are further improved by a special local search module based on systematic perturbations to the operation processing times. The superiority of the proposed approach is especially remarkable when the size of the scheduling problem is large.     

A Heuristic Search Approach Using Approximate Solutions to Petri Net State Equations for Scheduling Flexible Manufacturing Systems

This paper proposes a new heuristic search approach based on an analytic theory of the Petri net state equations for scheduling flexible manufacturing systems (FMSs) with the goal of minimizing makespan. The proposed method models an FMS using a timed Petri net and exploits approximate solutions of the net's state equation to predict the total cost (makespan) from the initial state through the current state to the goal. That is, the heuristic function considers global information provided by the state equation. This makes the method possible to obtain solutions better than those obtained using prior works (Lee and DiCesare, 1994a, 1994b) that consider only the current status or limited global information. In addition, to reduce memory requirement and thus to increase the efficiency of handling larger systems, the proposed scheduling algorithm contains a procedure to reduce the searched state space.     

Learning iterative dispatching rules for job shop scheduling with genetic programming

This study proposes a new type of dispatching rule for job shop scheduling problems. The novelty of these dispatching rules is that they can iteratively improve the schedules by utilising the information from completed schedules. While the quality of the schedule can be improved, the proposed iterative dispatching rules (IDRs) still maintain the easiness of implementation and low computational effort of the traditional dispatching rules. This feature makes them more attractive for large-scale manufacturing systems. A genetic programming (GP) method is developed in this paper to evolve IDRs for job shop scheduling problems. The results show that the proposed GP method is significantly better than the simple GP method for evolving composite dispatching rules. The evolved IDRs also show their superiority to the benchmark dispatching rules when tested on different problem instances with makespan and total weighted tardiness as the objectives. Different aspects of IDRs are also investigated and the insights from these analyses are used to enhance the performance of IDRs.     

基于Agent的芯片制造生产线动态调度方法研究

基于智能体技术,提出了芯片制造生产线动态调度新方法,实现了投料调度、工件调度与设备维护调度的集成。首先,给出了基于智能体的动态调度模型,该模型中包括管理智能体、投料智能体、工件智能体、设备智能体、运输智能体与人员智能体。投料智能体用于实现投料控制,工件调度通过工件智能体与设备智能体之间的协商实现,设备维护调度由设备智能体实现。通过智能体间的合作,能够实现投料控制、工件调度与设备维护调度的协同进行。然后给出了工件智能体与设备智能体问协商使用的协商协议（即带有时间约束的单步协商协议）,以及智能体决策中使用的调度算法（包括投料调度算法,工件调度算法与设备维护调度算法）。最后,通过实例,给出了方法的使用过程,并进行了模拟仿真。仿真结果表明,集成的调度方法能更好地优化模型的生产率、加工周期与设备利用率,最终提高准时交货率。     

An adaptive approach to dynamic scheduling in knowledgeable manufacturing cell

To overcome deficiency in the global capacity of a single dispatching rule, it is vital to select a dispatching rule in real time for dynamic scheduling. Among the studies addressing the method for selecting dispatching rules, few have no requirements for domain knowledge or accurate training example, which is hard to acquire from the real production system. In this paper, a new learning algorithm, along with the presentation of an adaptive scheduling control policy, is proposed to obtain the dynamic scheduling knowledge effectively, and different dispatching rules are selected to schedule the jobs in the machine buffer according to the current transient state of the system. Case studies are given to illustrate the validity of the scheduling control policy.     

A rolling horizon job shop rescheduling strategy in the dynamic environment

In this paper, the job shop scheduling problem in a dynamic environment is studied. Jobs arrive continuously, machines breakdown, machines are repaired and due dates of jobs may change during processing. Inspired by the rolling horizon optimisation method from predictive control technology, a periodic and event-driven rolling horizon scheduling strategy is presented and adapted to continuous processing in a changing environment. The scheduling algorithm is a hybrid of genetic algorithms and dispatching rules for solving the job shop scheduling problem with sequence-dependent set-up time and due date constraints. Simulation results show that the proposed strategy is more suitable for a dynamic job shop environment than the static scheduling strategy.     

基于零件虚拟工序队列的FMS动态调度研究

提出虚拟工序队列的概念,并在此基础上提出一种基于静态零件分批（静态调度）的ＦＭＳ动态生产调度方法,该算法充分考虑到ＦＭＳ的系统资源限制及零件运输等环节,能较好地解决生产调度中的设备负荷平衡,系统紧急事件发生（如机床故障,新零件加工等）,可替代加工工序等问题,且易于实现,具有较广泛的适应性。     

An effective immune algorithm based on novel dispatching rules for the flexible flow-shop scheduling problem with multiprocessor tasks

As a strongly NP-hard problem, the flexible flow-shop problem with multiprocessor tasks (FFSPMT) has gained much attention due to its academic significance and wide application background. To solve the FFSPMT, the dispatching rule is crucial to decode job order sequences to schedules, which has a great effect on the quality of the solution. In this paper, several novel dispatching rules are proposed to arrange the job processing order and machine assignment to minimize makespan of the FFSPMT by narrowing the idle time between the consecutive operations in the processor as well as by increasing the flexibility in selecting processors to schedule the following operations. With these rules, an immune algorithm (IA) is proposed to solve the FFSPMT, where special crossover, mutation, and vaccination operators are well designed and utilized. Meanwhile, some theoretical analysis for the local search operators is provided for guiding local search reasonably. The computational results based on 120 well-known benchmark instances and comparisons with some existing algorithms demonstrate the effectiveness of the proposed dispatching rules and the immune algorithm.     

考虑工序相关性的动态Job shop调度问题启发式算法

提出一类考虑工序相关性的、工件批量到达的动态Job shop 调度问题,在对工序相关性进行了定义和数学描述的基础上,进一步建立了动态Job shop 调度问题的优化模型。设计了一种组合式调度规则RAN(FCFS,ODD),并提出了基于规则的启发式算法以及该类动态Job shop 调度问题的算例生成方法。为验证算法和比较评估调度规则的性能,对算例采用文献提出的7种调度规则和RAN(FCFS,ODD)进行了仿真调度,对调度结果的分析表明了算法的有效性和RAN(FCFS,ODD)调度规则求解所提出的动态Job Shop 调度问题的优越性能。