循环取货带有时间窗约束的入库道口车辆调度

研究了循环取货模式下带有时间窗约束的入库道口车辆调度问题,为使车辆运输成本和取货时间成本、卸货时间成本最小,建立混合整数规划数学模型,设计了两阶段算法求解模型,第一阶段产生满足容量约束的较好初始解,第二阶段通过发车时间与路径同时编码的模拟退火算法进行求解,根据某汽车制造商循环取货的实际运作情况,构造算例并验证了该模型和算法的有效性。结论表明,制造商处的道口限制对循环取货发车时间与路径调度有较大影响,同时对发车时间和路径进行调度更有利于降低循环取货的运输费用。     

带准备时间的作业车间分批调度算法

研究了带有准备时间的作业车间分批排序问题,通过模拟计算将3种算法进行了对比分析.结果表明,遗传退火两阶段算法不受等分批限制,产生的各子批大小可以不同,这使得解空间更大,因而得到满意解的可能性大大增加.     

随机型双边混流装配线平衡问题的两阶段求解方法研究

为降低求解随机型双边混流装配线平衡问题的复杂性,提出了一种遗传算法与仿真分析相结合的两阶段求解方法。首先建立忽略装配线同一工作站组的两工作站之间作业先后顺序约束的随机型双边混流装配线平衡问题的简化数学模型,采用一种基于序列组合编码方式的遗传算法对简化模型进行求解,获取备选解;在此基础上,建立考虑所有约束条件的仿真模型,通过系统仿真分析与评价,从备选解中获得该问题的最优解(或次优解)。算例研究表明,所提出的两阶段求解方法,在获得满意解的同时,可以大幅度降低问题求解的复杂性。     

基于时间满意的覆盖问题及混合遗传算法实现

利用时间满意度函数,从顾客角度考虑覆盖半径,从企业角度考虑覆盖比例,提出了比传统集覆盖问题更一般的基于时间满意的覆盖选址问题.通过建立这一问题的整数规划模型,应用混合遗传算法对该问题求解,并同几种被证明为在覆盖选址问题中计算效果较好的启发式算法进行了对比实验分析,实验显示本文的算法策略可以得到高质量的解.     

带服务水平和容量约束的区域规划问题研究

综合考虑服务水平约束、养护车辆工作时长限制,建立了养护服务区域规划的混合整数规划模型,即选址弧路径规划模型。考虑对区域划分的要求,设计一个基于扫描算法的多项式时间混合启发式算法对问题进行求解。算法的第一阶段对各段道路进行聚类,第二阶段根据服务水平和惩罚成本将各个类分配给备选点,最后一个阶段对备选点进行合并,确定养护站选址及其负责的区域。经实验分析,混合启发式算法区域规划解总成本比分支切割算法高7.89%,其区域重叠度和紧凑度均好于分支切割算法。此外,考虑道路服务时间服从正态分布,对模型和算法进行改进。实验结果表明改进的算法能够以成本为代价降低超额工作量。     

危化品事故连续消耗型应急物资调度模型

应急物资调度是危化品事故应急救援工作的关键内容,危化品事故会对周边区域造成环境污染,其应急物资调度问题应该考虑多个目标。针对危化品事故应急物资调度特征,建立了连续消耗应急物资调度多目标模型。模型以缺货损失最小、调运时间最短和调运总环境风险最小为优化目标,考虑应急中心物资数量限制和救援时间限制等约束条件,采用NSGAⅡ遗传算法求解。对于求解结果,采用逼近理想解的排序方法进行选择。算例计算结果表明,多目标模型优化方案比单目标模型优化方案有较大的优势,同时说明算法合理可行。     

三维装载约束的车辆路径问题的模拟退火算法

传统的车辆路径问题只考虑物品装载的质量属性约束,而忽略其他装载属性约束.针对这种情况,研究了三维装载约束的车辆路径问题,提出了三维装载的处理算法,基于模拟退火算法设计了一种两阶段启发式算法进行求解:第1阶段通过启发式算法得到初始解;第2阶段通过模拟退火算法对初始解进行改进,构造了测试集对结果进行验证.实验结果显示该算法是能够有效的求解该问题.     

改进离散候鸟优化算法求解带缺失阶段的柔性流水车间问题

针对每阶段包含不相关并行机的柔性流水车间调度，研究了具有缺失阶段的总加权完工时间问题。由于该问题是NP-hard的，因此，提出基于两段式编码和组合邻域策略的改进离散候鸟优化算法进行求解。基于机器和工件编号设计两段式编码，利用最短加工时间规则和随机策略获得初始候鸟种群。领飞鸟和跟飞鸟进化中引入组合邻域策略以产生邻域解，最后对最差个体设计重置机制以再次提高解的质量。针对不同规模问题，对所提算法和四种启发式算法进行仿真实验，实验结果表明改进离散候鸟优化算法得到了更高质量的满意解。     

考虑班种约束的呼叫中心排班优化模型与算法

一些呼叫中心设置早班、晚班和两段班班种;排班时规定员工周内同班种的班次当班,并且限制两段班当班员工占比。构建该实际情景的坐席人员排班问题整数规划模型。鉴于问题难解性,首先通过问题结构层次分解,以及对班种与班次覆盖区段人力需求量化指标的表征刻画,提出启发式算法生成解方案;然后采用基于模拟退火机制的邻域搜索算法改善解方案。计算实验表明整数规划模型适于求解小规模排班问题最优解,而上述两阶段优化算法能够获得大规模问题优化解。研究表明,在优化人力成本情况下可兼顾坐席人员工作时间高规律性。     

应用层组播路由的一种改进神经网络解法

研究了使用双层递归神经网络(DRNN)模型求解应用层组播路由的问题.对原有模型的神经元矩阵及能量函数进行改变,并引入了新的线性编程神经元,解决了原模型不能求解组播路由的缺陷.与启发式组播路由算法相比,该解决方案的计算复杂度低,速度较快,而且与其它的神经网络相比,由于引入了基尔霍夫限制条件,保障了解的质量,且所使用的神经元数量少,而在解的精确度上则与其它算法相当.     

A three-phase heuristic algorithm for fixed-charge capacitated material flow network design with input/output points location

In this paper, we consider the material flow network design problem in which locations of input and output points of departments and flow paths are determined concurrently on a given block layout. The objective of the problem is to minimize the sum of transportation cost, flow paths construction cost and penalty cost for non-smooth material flows, i.e., flows with turns. A mixed integer programming model is given for the problem and a three-phase heuristic algorithm is developed to solve the problem. In the suggested algorithm, we generate an initial flow network by determining locations of input/output points and flow paths sequentially in the first and second phases, respectively, and then improve it by changing locations of input/output points and flow paths iteratively in the third phase. To evaluate the performance of the suggested algorithms, a series of computational experiments are performed on well-known problem instances as well as randomly generated test problems. Results of computational experiments show that the suggested algorithm gives good solutions in a short computation time.     

排列熵算法参数的优化确定方法研究

由于排列熵算法能够有效放大时间序列的微弱变化,且计算简单、实时性好,已在信号突变检测方面显示出良好的应用前景,但是排列熵算法中嵌入维数和延迟时间等参数的确定仍依赖于经验和尝试,该问题已成为排列熵算法走向工程应用的瓶颈问题。根据排列熵算法的原理,提出了基于重构时间序列最佳相空间来确定模型参数的方法。根据相空间重构的两种观点,介绍了延迟时间与嵌入维数独立确定和联合确定两种方法的基本理论,然后利用仿真信号和滚动轴承全寿命数据对两种算法进行了检验和对比。结果表明,模型参数的独立确定方法比联合确定方法对信号的异常检测更好。     

Space-time finite elements for a simplified thermo-metallurgical problem relevant to casting

The non-linear thermo-metallurgical problem, relevant for the cooling of a molten metal including the macro-segregation that occurs during the cooling process, is studied here. Due to the strong non-linearities involved in phase transformations, it is necessary to use a fine resolution in space-time in a finite element approximation in order to meet accuracy requirements. We derive space-time FE-methods based on the discontinuous and continuous Galerkin method in time for the energy equation. This formulation integrates the stored energy exactly for a given heat flux. When macro-segregation is incorporated into the model, the problem can be formulated in such a way that the phase-transition drives a flow of species. In addition, diffusion is possible throughout the domain. The model can be further rewritten using a potential approach. By this approach for modelling macro-segregation, we are able to obtain discretizations that guarantee that the balance equations are satisfied, and it is possible to solve the phase-transition problem either as a field problem or as a local problem (defined by a local evolution rule).     

A model of binary alloy solidification

A linear model for the solidification of a dilute binary alloy is presented. In this model the solidus and liquidus curves are linear. As a consequence internal energy depends linearly upon temperature and concentration. The formulation is a generalization of the well-known enthalpy method to treat a phase change problem involving coupled heat and mass transfer. Both analytic and numerical formulations are given. Results from the latter are presented and compared with an explicit solution of Rubinstein for a Stefan-like problem posed in a semi-infinite slab. Some remarks on the behaviour of the explicit solution are given.     

库存需求灾变灰预测模型

首先对控制库存需求常用的序列预测模型进行了回顾,指出它们存在的主要问题在于需要大量数据和模型设置的主观性等;接着提出了一类预测方法－－灰灾变预测模型以解决这些问题,并详细介绍了这类模型;然后给出一个案例进行相关的研究和分析;最后,对这类模型进行了总的评价。     

Analysis of parameter estimation heat conduction problems with phase change using the finite element method

A solution to a parameter estimation problem in heat conduction with phase change is presented. Specifically, the thermophysical properties of the material are estimated by using the temperature measurements of several sensors located inside the fixed domain. In the parameter estimation problem the error is minimized between given and modelled temperatures at sensor locations over the time interval where the experiment involved variation of the material parameters. The methods are illustrated with several examples. The effects on the solution of the locations of the sensors, the number of the sensors, the time step and the number of time steps are examined.     

Analysis and solution to the single-level batch production smoothing problem

Many companies use mixed-model production systems running under the Just-in-Time philosophy in order to efficiently meet customer demands for a variety of products. Such systems require demand be stable and production sequence be leveled. The production smoothing problem aims at finding level schedules in which the appearances of products are dispersed over the horizon as uniformly as possible. In this paper, the production smoothing problem is extended to a more general manufacturing environment where a single machine can be identified as either the final or the bottleneck stage of the system and products may have arbitrary non-zero setup and processing time requirements on this single machine. An optimization model is built for the problem and a two phase solution methodology is developed. The first phase problem is shown to be NP-hard and a parametric heuristic procedure is proposed for its solution. In contrast, the second phase problem is shown to be efficiently solvable and currently available solution methods are adopted from the literature. A computational study is designed to test the proposed two phase solution methodology and also the parametric heuristic procedure. Computational results show that the proposed two phase solution methodology enables effective and efficient control of the studied manufacturing system, and the heuristic procedure developed for the first phase problem is time efficient and promises near optimal solutions for a variety of test instances.     

Layout planning for facilities with fixed shapes and input and output points

This paper focuses on the facility layout problem in which each facility has a predetermined shape and input and output points. In the problem, facilities are placed within a given floor, and the spatial coordinates and orientation of each facility are to be determined. We give a mixed integer programming (MIP) model for the layout planning problem with the objective of minimizing the sum of rectilinear distances weighted by flow amounts between input and output points of the facilities. Using the MIP model, we develop a two-phase algorithm in which an initial layout is generated in the construction phase and is improved using four improvement methods applied iteratively in the improvement phase. Results of computational experiments show that the proposed algorithm gives better solutions than existing algorithms.     

圆染色在城市交叉路口交通信号相位个数设计中的应用

以城市交叉路口交通信号相位的优化为背景,分析了用圆染色解决这个问题的合理性,并给出现实中了几类交叉路口的最优相位个数。