随机抽样算法在多项目调度中的应用

针对资源约束下的多项目调度问题,扩展了单项目进度生产机制,并在此基础上设计了多项目调度的随机抽样算法。通过对比分析,说明不同的任务优先规则对随机抽样算法具有不同的影响,其中采用MINSLK等优先规则的随机抽样算法能够有效地缩短项目平均工期。与其他多项目调度启发式算法的比较和统计检验,说明随机抽样算法显著优于这些常用的启发式算法。     

资源约束下多项目调度的改进遗传算法

针对资源约束下的多项目调度问题,在前人提出的有效的启发式算法研究路径基础上,本文利用遗传算法,结合进度生成机制,提出了多项目调度的改进遗传算法。与其他多项目调度启发式算法相比,该算法在平均项目延迟和最佳解比例方面都表现较好,综合利用优化后的优先规则也使得该算法更适用于不同网络复杂度和不同资源约束程度的多项目调度问题中。     

航空票务公司免费接送服务中的车次调度模型及算法

本文从顾客满意度和成本出发，对航空票务公司免费接送顾客的车次分配和调度问题进行研究，首先分析了票务公司免费接送服务的业务流程；然后给出了基于时间的顾客满意度函数，并建立了在一定顾客满意度下成本最小的车次调度模型；最后利用集划分方法求解该模型。集划分包括以下两步：第一步采用广度优先搜索生成全部车次集；第二步提出了一种启发式算法。通过仿真分析比较，验证了启发式算法的有效性。     

面向计算机集群系统的FP-Growth算法的并行计算

FP-Gfowlth是频繁模式挖掘的经典算法,能够在不产生候选集的情况下生成所有的频繁模式,效率与Apri-ori算法相比有巨大提高,然而FP-Growth算法在挖掘频繁模式过程中需要递归构建大量的条件FP-tree,并分别针对这些条件FP-tree进行挖掘,时间及空间效率不高,在实际应用中存在很大局限性.计算机集群是由多台普通计算机设备通过特定方式结合在一起构成的并行处理系统,属于分布式计算环境,具有计算能力强大、性价比高、灵活等优势.本文提出一种面向计算机集群的并行挖掘算法Gridify FP-Growth,该算法以FP-Growth为基础,通过任务划分的形式.将计算任务分配到计算机集群中各个计算节点上执行,充分利用各个节点的计算资源,最后汇总各节点的计算结果.实验证明Gridify FP-Growth算法不会牺牲计算的准确性,并可以大幅度缩短计算时间,有效缓解计算大规模数据库时的内存压力.     

基于鲁棒优化模型的项目调度策略遗传算法

针对任务工期不确定的资源受限项目调度问题,提出了求解该问题的鲁棒优化数学模型。在随机规划基础上,针对该鲁棒优化模型设计了遗传算法。算法通过多种任务优先规则生成初始种群以保证种群多样性。应用该遗传算法对项目实例进行了求解,分析表明所求得的项目调度策略能够有效应对任务工期不确定性导致的随机差异,具有较强的鲁棒性。参数分析表明决策者可以通过调节模型权重系数有效平衡解的可行性与最优性,有助于决策者根据风险偏好进行选择。     

山区配电网震后恢复及资源调度集成优化研究

地震的发生往往会对配电网产生较大破坏，影响供电质量。山区路网地形地貌的复杂性及受地震破坏等，增加了资源调度的困难，阻碍了山区配电网的恢复。因此，本文针对山区配电网震后恢复及资源调度集成优化问题，建立以恢复绩效最大为目标的非线性混合整数规划模型，并根据模型特点，引入A*算法求解各节点间的最优路径与所需时间，通过改进具有快速稳健特性的细菌群趋药性算法（BCCOA）对模型进行求解，得到山区配电网恢复及资源调度策略，并采用IEEE30，IEEE57和IEEE118节点系统，验证所述策略的可行性和有效性。结果表明：（1）改进的BCCOA相对于枚举法，在求解精度和计算时间上都有较好优势；（2）当资源调度能力一定，改变抢修能力时，配电网恢复绩效平均提升36.40%。而当抢修能力一定，改变资源调度能力时，由于受抢修能力的影响，配电网恢复绩效平均提升7.99%，但资源调度延迟率下降了61.71%，可提高配电网恢复绩效。而同时改变两者的能力时，配电网恢复绩效平均提升38.23%，提升了209.14%，表明资源调度在山区配电网震后恢复中的重要性，说明山区配电网震后恢复及资源调度集成优化研究符合实际决策需求；（3）改进的BCCOA相较于BCCOA，其算法性能提升了2.68%，各改进策略均产生了效果。     

活动重叠模式与资源约束下的项目调度优化

采取活动重叠模式通常是加速研发的有效手段,带有活动重叠的资源受限项目调度问题是经典资源受限项目调度问题的扩展.首先,深入分析了活动重叠对于项目调度的影响,对活动重叠及其不确定进行详细描述与建模,提出了活动重叠导致下游活动返工时间的二项分布概率模型;其次,构建了以最小化研发项目期望工期为目标的优化调度模型,设计了基于串行进度生成机制的遗传算法对大规模问题进行优化求解;最后,基于PSPLIB J60问题库中480个算例分析了该算法的计算结果,并考察了网络参数、资源参数和重叠参数变化时,采用活动重叠模式对缩短项目工期的影响.研究结果表明:活动对资源的需求强度越小或资源稀缺程度越低,可重叠活动对数量就会增加,项目工期缩短得越明显;网络复杂度的变化对缩短项目工期的影响不大;项目中重叠活动对越多,重叠导致的下游活动返工的概率越小,项目工期缩短的越明显.     

时间窗口约束下基于改进蚁群算法的协同制造调度研究

与传统调度模式不同，协同制造模式下企业之间的调度模式极其复杂。协同企业间的加工工序路线并不固定，且不同类型产品具有不同的加工路线网络。为此本文针对平衡型、瓶颈型、跳跃型、混合型四类具有典型特点的协同制造网络G_p进行分析和设计；考虑制造企业同类产品合并加工策略，构建基于连续加工量的分段生产成本函数；通过设计合理的订单最早交货时间和最晚交货时间，对订单交货进行时间窗口约束，并在此基础上构建了由制造商生产成本W_c^m、订单等待W_s^k（Q_k， T'_k）和提前完工库存成本W_s^k（Q_k， T″_k）、延期惩罚成本构成W_l^k（Q_k， T'″_k）的目标函数。为求解该模型，创新性将蒙特卡洛思想引入蚁群算法，提高蚂蚁选择合理性，避免局部最优；同时，采用移动窗口[min， max]奖励机制，并且对信息素奖励乘以平衡系数k（N）提高奖励可信度，加快搜索速度并提高求解性能。仿真结果表明，本文构建调度模型合理，可以获得优化的调度结果；同时，本文提出的蚁群改进寻优算法具有良好的求解速度和收敛性，算法具有较好的稳定性。     

复杂不确定环境下活动可拆分的项目资源鲁棒性调度优化

复杂不确定环境下,制定一个具有较强抗干扰能力的基准进度计划非常必要。本文研究了活动工期不确定环境下考虑活动可拆分的项目资源鲁棒性调度优化问题,旨在考虑活动可拆分,探究在活动优先关系约束、项目截止日期约束、活动拆分约束、资源流约束等条件下如何进行活动拆分决策并合理地安排各个项目活动/活动分段间的资源调配方案和时间缓冲添加策略,以制定鲁棒性最大化的基准进度计划。本文创新点如下：1)在项目资源鲁棒性调度优化问题中考虑活动可拆分,定义了资源流网络下活动自由时差的计算方法,提出了一种新的活动可拆分情形下进度计划鲁棒性的衡量方式,进而构建得到了考虑活动可拆分的项目资源鲁棒性调度优化模型;2)分析证明了问题模型的强NP-hard属性以及非线性属性,并在此基础上开发了一种内嵌资源调度安排的遗传算法进行问题求解;3)选取一个典型的实际案例对研究问题进行说明,直观展示了活动拆分执行对进度计划鲁棒性提升的重要价值,揭示了鲁棒性调度计划中资源调度方案的重要性,得到了活动拆分执行会增加项目内部资源转移次数的结论。     

考虑交通时间与设备转换时间的跨区域就医调度优化问题

在资源共享时代背景下,跨区域就医可以很好地解决患者日益增长的就医需求与医疗资源紧张的矛盾。本论文以医疗联盟为研究对象,在关键医疗资源共享的前提下,通过患者跨区域就医实现就医诊断延误最小化,以满足患者就医需求。本研究同时考虑了患者跨区域交通时间与基于患者诊断类型的设备转换时间,以最小化患者就医总延迟为目标,分配患者就诊医院及优化患者就诊/检查顺序。针对该问题,论文首次提出以最早交货期原则（EDD rule）为基础,以患者再分配为主导的EDD-ReAss1和EDD-ReAss2启发式算法,结合局部搜索算法以进一步提高就医调度方案的质量,缩短患者诊断/检查等待时间。实验结果表明,新启发式算法EDD-ReAss1和EDD-ReAss2算法性能显著好于EDD,SPT和LPT等调度规则;在较短运算时间内Swap局部搜索算法性能最优。     

A heuristic approach to n/ m job shop scheduling: fuzzy dynamic scheduling algorithms

n/m shop scheduling is a ‘ NP-Hard’ problem. Using conventional heuristic algorithms ( priority rules) only, it is almost impossible to achieve an optimal solution. Research has been carried out to improve the heuristic algorithms to give a near-optimal solution. This paper advocates a fuzzy logic based, dynamic scheduling algoridim aimed at achieving this goal. The concept of new membership functions is discussed in die algorithm as a link to connect several priority rules. The constraints to determine the membership function of jobs for a particular priority rule are established, and three membership functions are developed. In order to decide the weight vector of priority rules, an aggregate performance measure is suggested. The methodology for constructing the weight vector is discussed in detail. Experiments have been carried out using a simulation technique to validate the proposed scheduling algorithm.     

Release of maintenance jobs in a decentralized multi-stage production/maintenance system with continuous condition monitoring

Condition-based maintenance is analyzed for multi-stage production with a separate maintenance department. It is assumed that the conditions of the machines deteriorate as a function of multiple production parameters and that the task of maintenance is to keep up predefined operational availabilities of the individual machines. In this context the problem of determining the optimal machine condition that triggers the release of a preventive maintenance job and the problem of scheduling maintenance jobs at the maintenance department arise. Existing approaches to solve these problems either assume a monolithic production/maintenance system or concentrate on a decentralized system in which the information flow and resource transfer do not cause delays. With our paper we aim at (1) deriving a triggering mechanism that is able to cope with relaxed assumptions and at (2) developing specific priority rules for scheduling maintenance jobs. Therefore, in this paper a specific continuous condition monitoring and a suitable information exchange protocol are developed, factors determining the release situation are operationalized, impacts of choosing the triggering conditions are identified and the components of specific priority rules for scheduling maintenance jobs are clearly elaborated. Finally the performance of the resulting solution approach is analyzed by simulations. Thereby, relevant characteristics of the production/maintenance system, the maintenance task and relevant priority rules are varied systematically. This research contributes answers to the questions on how the exchange of local information can be structured, the parameters of condition-based maintenance can be set and on what maintenance-specific priority rules can be applied in case of incomplete information about deterioration in a decentralized multistage production/maintenance system.     

一种求解双目标flow shop排序问题的进化算法

提出一种求解双目标flow shop排序的递进多目标进化算法.算法采用改进的精英复制策略,在实现精英保留的前提下降低了计算复杂性;通过递进进化模式增加群体多样性,改善了算法收敛性;通过群体进化过程中对非劣解集进行竞争型可变邻域启发式搜索,增强了算法局部搜索性能.采用新算法和参照算法NSGA-II对31个标准双目标flow shop算例进行优化.研究结果表明,新算法在所有算例的求解中均获得了优于NSGA-II的非劣解集,验证了算法的有效性.     

基于拍卖机制的资源转移时间型动态分布式多项目调度

实践中,企业并行实施地域上分散的多个项目时,资源在各子项目之间的转移时间是影响多项目整体进度的关键因素,同时在动态多项目环境下,新项目不断到达且到达时间不可预知使得制定多项目调度计划遭遇更大困难。本文在动态环境下对资源转移时间型分布式多项目调度问题进行建模和求解,基于多代理系统建立分布式多项目调度问题的动态模型,并将拍卖理论引入其中,设计一种基于时间窗拍卖机制的分布式多代理系统(DMAS/ATW),在动态环境和资源转移时间约束下为多项目配置全局资源。通过一个具体的分布式多项目示例详细分析DMAS/ATW算法的动态调度过程,并基于MPSPLIB中的分布式多项目算例开展数值实验。实验结果表明：无资源转移时间约束时,DMAS/ATW算法求得的平均项目延迟同比相关文献中的DMAS/RIA算法最多减少42%,平均减少26%;有资源转移时间约束时,DMAS/ATW算法对1/3算例集的求解结果优于DMAS/RIA算法在无资源转移时间约束时的结果,验证了本文DMAS/ATW算法求解效果的优异性。对算例规模和全局资源利用系数的实验分析还表明,DMAS/ATW算法对不同规模和资源约束紧张程度的算例都具有良好的适应性。     

The Effects of Learning,Forgetting, and Relearning on Decision Rule Performance in Multiproject Scheduling

Product development occurs in multiproject environments where preemption is often allowed so that critical projects can be addressed immediately. Because product development is characterized by time-based competition, there is pressure to make decisions quickly using heuristics methods that yield fast project completion. Preemption heuristics are needed both to choose activities for preemption and then to determine which resources to use to restart preempted activities. Past research involving preemption has ignored any completion time penalty due to the forgetting experienced by project personnel during preemption and the resulting relearning time required to regain lost proficiency. The purpose of this research is to determine the impact of learning, forgetting, and relearning (LFR) on project completion time when preemption is allowed. We present a model for the LFR cycle in multiproject development environments. We test a number of priority rules for activity scheduling, activity preemption, and resource assignment subsequent to preemption, subject to the existence of the LFR cycle, for which a single type of knowledge worker resource is assigned among multiple projects. The results of the simulation experiments clearly demonstrate that LFR effects are significant. The tests of different scheduling, preemption, and resource reassignment rules show that the choice of rule is crucial in mitigating the completion time penalty effects of the LFR cycle, while maintaining high levels of resource utilization. Specifically, the worst performing rules tested for each performance measure are those that attempt to maintain high resource utilization. The best performing rules are based on activity criticality and resource learning.     

Multimode time-cost-robustness trade-off project scheduling problem under uncertainty

The time/cost trade-off problem is a well-known project scheduling problem that has been extensively studied. In recent years, many researchers have begun to focus on project scheduling problems under uncertainty to cope with uncertain factors, such as resource idleness, high inventory, and missing deadlines. To reduce the disturbance from uncertain factors, the aim of robust scheduling is to generate schedules with time buffers or resource buffers, which are capped by project makespan and project cost. This paper addresses a time-cost-robustness trade-off project scheduling problem with multiple activity execution modes under uncertainty. A multiobjective optimization model with three objectives (makespan minimization, cost minimization, and robustness maximization) is constructed and three propositions are proposed. An epsilon-constraint method-based genetic algorithm along with three improvement measures is designed to solve this NP-hard problem and to develop Pareto schedule sets, and a large-scale computational experiment on a randomly generated dataset is performed to validate the effectiveness of the proposed algorithm and the improvement measures. The final sensitivity analysis of three key parameters shows their distinctive influences on the three objectives, according to which several suggestions are given to project managers on the effective measures to improve the three objectives.