敏捷供应链静态调度模型及其贪婪算法

针对确定性市场需求下的敏捷供应链(ASC)调度具有批量和时间双重约束的问题,根据供应层级关系建立多级供应链结构模型(SCSM),在此基础上设计调度问题的线性规划(LP)模型及其两阶段求解算法：第一阶段的最短响应时间调度贪婪算法判定供应链系统对需求批量和时间约束满足与否;第二阶段的精益调度贪婪算法求解以最小化供应链库存成本为目标的最优调度方案。最后通过算例验证了模型和算法的实用性和有效性。     

效率异质型员工项目调度算法研究

研究了员工具有异质效率、最小化项目工期的项目调度问题,并建立了相应的整数线性规划模型。为解决此NP-hard问题,提出了基于优先规则的启发式算法,其在每次迭代中根据优先约束和优先规则选择优先任务员工对以分配任务,直至所有任务都完成调度。通过应用启发式算法生成初始调度,选用交换邻域结构和插入邻域结构产生邻域调度,并使用改进的前向递归算法求解目标函数值,构造出混合模拟退火算法。数值实验显示该算法能快速准确地进行寻优。     

网络集群计算系统中的并行任务调度

基于多处理机并行任务调度模型，探讨网络集群计算系统中的并行任务调度问题，首先证明了一般网络集群计算系统中调度算法的可近似性难度，然后提出了三种不同的启发式算法：最大长度优先调度算法、最大宽度优先调度算法和最大面积优先调度算法；然后根据大量的模拟实验对这些算法以及文献中已提出的调度算法进行了比较分析，结果表明该文的启发式算法比文献中的算法在性能上效果更好。     

基于有向无环图网格独立任务调度算法*

提出了基于有向无环图多约束网格环境下独立任务的调度模型,为其建立多约束线性规划模型,通过求解模型节点的优先级,获得网格各计算节点最优任务调度数;然后基于多约束最优任务调度方案,提出多约束带宽优先启发式算法(MCOPBHATS)和多约束计算速度优先启发式算法(MCOPCHATS)。实验结果表明,在多约束异构的网格环境下实现大量独立任务调度时, MCOPBHATS和MCOPCHATS算法的性能优于基于多约束最优任务调度方案的MinMin 算法。     

一种启发式与/或优先约束任务调度算法

系统描述了与或网模型及与或优先约束任务调度的可行性判定算法.以顶点覆盖问题为基础,证明与或优先约束任务调度最小完成时间问题是NP完全的.提出一种启发式调度算法,解决与或优先约束任务调度最小完成时间问题.仿真结果表明,该算法在降低算法复杂度的同时较其它相关算法具有更好的调度性能,从而证明在实时优先约束任务调度中引入图优化的理论是解决优先约束任务调度问题的一个有效途径.     

基于树型Petri网的网格资源调度模型

讨论树型网格环境下独立任务的调度问题,把网格资源计算能力和网络通信速度的异构资源调度问题转化为线性规划问题。提出网格资源调度的树型Petri网模型,基于速度优先的启发式资源调度算法对树型Petri网模型进行性能分析,验证了树型Petri网的单口模式传输特点。     

机器人反向动力学方程的并行计算

以PUMA560机器人的分解牛顿-欧拉反向动力学方程为模型,提出了方程分解的原则, 由此得到AOE(Activity On Edge)有向图.以此为基础,按照深度和时差的概念建立了 L-W优先表,并导出了一种启发式的调度算法.该算法在微处理机个数一定的情况下,可得 到最小调度时间.最后,以Stanford机器人的递推牛顿-欧拉反向动力学方程为例,说明了 该算法的有效性.     

三峡工程两坝联合通航调度的混合模拟退火算法

对三峡大坝和葛洲坝的一共5座船闸进行统一的船舶通航调度管理,是提高长江三峡水域航运能力的关键,然而其优化调度算法还缺乏必要的研究.本文首先提出了该问题的混合整数非线性规划模型,在实际通航调度环境中,该模型属于强NP-hard复杂度的大规模组合优化问题,因此设计了一种混合模拟退火算法来搜索次优化调度方案,该算法将解分解为闸次时间表和船舶调度计划两部分,在搜索过程中用启发式规则对闸次时日表进行调整,然后用深度优先搜索(DFS)算法根据闸次时间表求解船舶调度计划,最后根据Metropolis规则对当前解进行更新.针对实际通航数据的测试结果表明其优化效果明显优于原有的启发式算法.目前该算法已经成功地应用于实际的两坝联合通航调度系统中.     

考虑能耗成本和拖期成本的非同等并行机调度

针对橡胶轮胎硫化车间能源消耗大,浪费严重的现象,提出一类考虑能耗成本与拖期成本的非等同并行机调度问题,建立基于硫化机正常运行、空闲、停机三种运行状态的能源消耗成本与拖期成本的调度模型。设计了基于优先调度规则的启发式算法、基于能耗优化的启发式算法、组合启发式算法用于模型求解,并通过仿真实验分析、比较了各种算法的有效性与适用环境。同时,仿真实验结果也表明本文提出的考虑能耗成本和拖期成本的非同等并行机调度问题具有一定的理论与实践意义。     

虚拟计算环境中的多机群协同调度算法

基于虚拟计算环境的核心机理,提出由自主调度单元、域调度共同体、元调度执行体为核心的多机群协同系统框架.剖析多机群任务并发运行性能模型,设计了多机群协同调度算法框架,提出最大空闲节点优先、最小网络拥塞优先、最小异构因子优先与最小异构空闲节点优先4种启发式资源选择策略.实验验证了协同调度模型与算法在任务集完成时间与系统平均利用率的测度上的有效性.     

改进的哈里斯鹰算法求解农产品供应链调度

针对农产品流通体系的效率低、流通链条协同效率不高、紧急情况下食品供给慢等问题,通过将农产品供应链调度问题建模成混合流水车间调度问题。结合禁忌搜索算法中禁忌表机制,离散化实数编码,提出了一种改进的哈里斯鹰算法来求解农产品供应链调度问题。该方法相比较原始的哈里斯鹰算法,降低了算法陷入局部最优的可能,进一步提高了算法的求解精度。实验结果表明：相比较对比算法,改进的哈里斯鹰算法在提出的农产品供应链调度问题模型上取得了更好的效果。     

基于蚁群算法的MC供应链调度优化研究

为解决大规模定制模式下客户订单分离点后的动态供应链调度问题，提出了包括供应商选择及企业合作时序安排的优化调度模型，设计了基于蚁群算法的求解过程。通过多组数据实验及结果比较分析，对模型算法的有效性、稳定性进行了验证。     

Partner selection model and soft computing approach for dynamic alliance of enterprises

Partner selection is an active research topic in agile manufacturing and supply chain management. In this paper, the problem is described by a 0-1 integer programming with non-analytical objective function. Then, the solution space is reduced by defining the inefficient candidate. By using the fuzzy rule quantification method, a fuzzy logic based decision making approach for the project scheduling is proposed. We then develop a fuzzy decision embedded genetic algorithm. We compare the algorithm with tranditional methods. The results show that the suggested approach can quickly achieve optimal solution for large size problems with high probability. The approach was applied to the partner selection problem of a coal fire power station construction project. The satisfactory results have been achieved.     

Solving the integrated scheduling of production and rail transportation problem by Keshtel algorithm

Nowadays, scheduling of production cannot be done in isolation from scheduling of transportation since a coordinated solution to the integrated problem may improve the performance of the whole supply chain. In this paper, because of the widely used of rail transportation in supply chain, we develop the integrated scheduling of production and rail transportation. The problem is to determine both production schedule and rail transportation allocation of orders to optimize customer service at minimum total cost. In addition, we utilize some procedures and heuristics to encode the model in order to address it by two capable metaheuristics: Genetic algorithm (GA), and recently developed one, Keshtel algorithm (KA). Latter is firstly used for a mathematical model in supply chain literature. Besides, Taguchi experimental design method is utilized to set and estimate the proper values of the algorithms’ parameters to improve their performance. For the purpose of performance evaluation of the proposed algorithms, various problem sizes are employed and the computational results of the algorithms are compared with each other. Finally, we investigate the impacts of the rise in the problem size on the performance of our algorithms.     

A constraint programming model for the scheduling of JIT cross-docking systems with preemption

In this paper, a scheduling problem of minimizing the total of the earliness, tardiness and the number of preemption for the outbound trucks on a cross-dock system is considered. This problem, which is known to be NP-hard, is compatible with the concepts of just-in-time (JIT) production and supply chain management. A new multi-criteria model, with non-linear terms and integer variables, which cannot be solved efficiently for large sized problems, is proposed. This paper also shows how to map a JIT cross-dock model to a constraint satisfaction problem (CSP) and integer programming (IP). To solve the model for real size applications, a genetic algorithm (GA) is applied. Finally, a computational experiment is carried out to analyze the performances of CSP, GA and IP models with respect to modeling capability, solution quality and time.     

混合粒子群算法和多Agent系统的协同生产调度及其组合优化

为了从整体角度优化调度供应链网络的各个环节,研究了在供应链环境下核心制造商与制造商的协同生产调度方案,考虑在满足产品生产时间节点的最少加工时间。建立了多Agent的供应链环境下的协同生产调度模型,针对此模型设计了协同混合粒子群优化算法并进行求解。通过实例研究表明,供应链环境下制造商的协同优化对生产的计划与实行起到了关键的作用。     

基于调度规则和免疫算法的作业车间多目标调度

利用动态在线调度方法对动态环境下的作业车间进行研究,采用优先级调度规则对大量调度案例进行求解,针对7个调度目标,从备选调度规则集中选出了单个目标下性能最优的调度规则;为实现调度规则的动态选择以适应多目标调度,基于免疫系统中的独特型网络理论,设计了一种免疫调度算法．根据算法,定义了有效的抗体和抗原结构,并通过抗体间亲和力计算、抗体浓度计算、抗体选择等关键步骤,实现对调度规则的动态控制．仿真测试数据表明,所设计的免疫调度算法能根据不同的车间情况,快速选出不同的调度规则满足多个调度目标,有效解决了作业车间多目标调度问题．     

逆向物流供应链最佳供应商选择及订单量分配*

为解决逆向物流供应链中,供应商选择、订单量分配和提货点位置等不确定问题,建立了一个新的模糊多目标数学模型来确定最佳供应商选择、供应量及提货点位置,为避免在解决多目标模型时人为主观赋权,运用基于模糊目标规划的蒙特卡罗仿真模型来求解帕累托(pareto)理想解,采用遗传算法进行求解,并给出了相应优化方案,在此基础上研究讨论了不同权重分配下结果的优劣性及供应商选择风险,最后,针对不同权重分配,比较了遗传算法和Gurobi求解,实验表明,对于该问题模型遗传算法在解的优劣性上优于Gurobi。     

基于Web服务的敏捷供应链系统

如何根据敏捷供应链的特点,快速构造动态、异构的供应链系统是供应链研究中亟待解决的一个重要问题。文章通过分析现有的供应链集成技术的缺陷和Web服务的优点,建立了一个基于Web服务的敏捷供应链模型,并加以实现。     

Dynamic supply chain scheduling

Based on a combination of fundamental results of modern optimal program control theory and operations research, an original approach to supply chain scheduling is developed in order to answer the challenges of dynamics, uncertainty, and adaptivity. Both supply chain schedule generation and execution control are represented as an optimal program control problem in combination with mathematical programming and interpreted as a dynamic process of operations control within an adaptive framework. Hence, the problems and models of planning, scheduling, and adaptation can be consistently integrated on a unified mathematical axiomatic of modern control theory. In addition, operations control and flow control models are integrated and applicable for both discrete and continuous processes. The application of optimal control for supply chain scheduling becomes possible by formulating the scheduling model as a linear non-stationary finite-dimensional controlled differential system with the convex area of admissible control and a reconfigurable structure. For this model class, theorems of optimal control existence can be used regarding supply chain scheduling. The essential structural property of this model are the linear right parts of differential equations. This allows applying methods of discrete optimization for optimal control calculation. The calculation procedure is based on applying Pontryagin’s maximum principle and the resulting essential reduction of problem dimensionality that is under solution at each instant of time. The gained insights contribute to supply chain scheduling theory, providing advanced insights into dynamics of the whole supply chains (and not any dyadic relations in them) and transition from a partial “one-way” schedule optimization to the feedback loop-based dynamic and adaptive supply chain planning and scheduling.