蚂蚁算法在复杂性运输路径问题中的应用——多车场多车型路径问题

车辆路径问题中,行驶路线往往取决于一系列约束条件,如配送中心个数,货物需求量,交发货时间,车辆容量限制等。要想达到一定的目标,如路程最短,费用最小,时间尽量少,车辆尽量少等,就得借劲于合适的算法去解决实际的问题。蚂蚁算法在解决著名的旅行商（TSP）问题上已取得了很好的成效,目前已陆续渗透到其他问题的求解上。文章主要针对多车场多车型车辆路径问题,用蚁群算法以及蚁群算法的优化算法去解决一些实际问题。     

Vehicle flow formulation for two-echelon time-constrained vehicle routing problem

Two-echelon routing problems, including variants such as the two-echelon vehicle routing problem (2E-VRP) and the two-echelon location routing problem (2E-LRP), involve assignment and location decisions. However, the two-echelon time-constrained vehicle routing problem (2E-TVRP) that caters to from-linehaul-to-delivery practices does not involve assignment decisions. This routing problem variant for networks with two echelons has not yet attracted enough research interest. Localized or long-distance services suffer from the lack of the assignment decisions between satellites and customers. Therefore, the 2E-TVRP, rather than using assignment decisions, adopts time constraints to decide the routes on each of the two interacting echelons: large-capacity vehicles transport cargoes among satellites on the first echelon, and small-capacity vehicles deliver cargoes from satellites to customers on the second echelon. This study introduces a mixed integer linear programming model for the 2E-TVRP and proposes a heuristic algorithm that incorporates the savings algorithm followed by a variable neighborhood search phase. Illustrative examples are used to test the mathematical formulation and the heuristic and a case study is used to demonstrate that the heuristic can effectively solve realistic-size instances of the 2E-TVRP.     

微粒群算法在现代物流系统优化中的应用

根据物流系统优化问题和微粒群优化算法的各自特点，提出了采用微粒群优化算法求解物流系统优化问题的相关策略，对微粒群优化算法在旅行商问题、车辆路径问题、配送中心选址问题、布局优化问题上的应用进行了研究。     

基于扫描算法的启发式构造方法比较

针对车辆路径问题,基于扫描算法第一阶段的解,应用启发式算法中的最近插入算法、凸包算法和最远插入算法求解第二阶段。通过仿真实验,从总里程和算法运行时间两个方面对各算法性能给出评价。结果显示,在应用扫描算法进行聚类后,求解路径排程阶段,凸包算法虽然用时多于其它两种算法,但在里程上有明显优势,最远插入算法与最近插入算法在运行时间上没有显著差别,但在总里程上,前者较好。     

基于免疫—蚁群算法的应急物流配送网络优化研究

论文针对应急物流的配送车辆路径优化问题,建立以应急物流初期有限资源的分配达到整体的最优化（时间短、效率高、安全高、成本低）的多目标模型,结合模型的特点,利用免疫算法和蚁群算法相融合的启发式算法进行求解,最后进行算例分析,解决应急物流配送车辆路径选择的实际问题。     

A parallel insertion heuristic for vehicle routing with side constraints

In the early eighties, Fisher and Jaikumar developed a generalized assignment heuristic for vehicle routing problems. In this paper, we discuss some of the strong and weak points of this heuristic, and take its basic ideas to develop a new parallel insertion heuristic for the vehicle routing and scheduling problem that is better able to handle various side constraints.     

基于蚁群系统算法的车辆路径问题研究

通过引入改进的蚂蚁算法--蚁群系统算法来解决基本的车辆路径问题。针对具体实例运用计算机编程得出了基于蚁群系统算法的最优解,并与其它两种启发式算法进行对比,指出了蚁群系统算法的优势与存在的不足,提出了进一步研究的方向。     

现代物流中车辆路径问题的研究

本文阐述了车辆路径问题(VRP)的几种典型的模型和常见算法，分析了这些算法用于计算车辆路径问题的优缺点，并指出了VRP的研究范围还需向范围更广的供应链方向扩展。     

基于人员和车辆可租赁条件下的车辆路径选配模型及优化

物流服务供应商面临的决策问题常常是复杂的规划问题。文中考虑短程的计划,物流服务供应商在需求点的时间窗、车辆的容积和人员的调度,包括租用的人员和车辆的条件下决定收发件路线,建立了路线和人员调度相结合的规划模型。采用节约启发式算法来求解车辆路径问题,得到路线安排的满意解。之后采用贪婪启发式算法对人员进行分配,通过实际案例得出:在租赁人员和车辆的情况下比不租赁人员和车辆总的费用更低。结果表明,启发式算法既能够满足时间限制,又能够节约运输里程和费用。     

多旅行商问题研究综述

多旅行商问题是经典旅行商问题的一种泛化,加上某些特定的附加条件,则能够演化成一些较现实的问题,因而具有较高的理论研究和应用价值。在多旅行商问题中,一个任务由多位旅行商共同完成,其问题的求解难度较经典旅行商问题更大,用于经典旅行商问题求解的方法或策略不能简单地应用于多旅行商问题的求解,有关该问题的研究成果远比经典旅行商问题少,本文针对目前有关多旅行商问题的研究进展情况进行综述。     

基于PDDL智能规划的车辆路径问题研究

针对车辆路径问题和多种扩展问题,文章提出一种基于PDDL智能规划的建模和求解方法,给出了多种车辆路径问题的PDDL模型,并利用SGPlan6和DaeYa规划器对PDDL模型进行了求解。实验结果表明,该算法能够在较短时间内给出高质量的车辆路径安排方案,并具备很好的扩展能力。     

Managing returnable transport items in a vendor managed inventory system

This paper addresses a closed-loop inventory routing problem with demand uncertainty, which manages the delivery operations to customers and pick-up operations of empty returnable transport items (RTIs) from customers. The problem involves decisions regarding forward and backward vehicle routes, the delivery and collection quantities, the amount of production in terms of the number of filled RTIs, and the number of RTIs produced/ bought by the vendor during the defined planning horizon. The problem considers the holding costs, fixed cost of operating vehicles, fuel consumption cost, producing/buying costs, cleaning costs of RTIs, and loading and unloading costs. We formulate the problem as a mixed integer linear programming model and propose a Relax and Fix based solution approach to solve large instances. We conduct extensive analyses on a case study derived from a fruit and vegetable distribution network and several hypothetical instances. Our analyses investigate the effects of several problem parameter changes (i.e., average collection rate, customer service level, cost of buying crate and handling cost) on the total logistics cost. Additional numerical analyses are performed to demonstrate the usage of the model for evaluating the cost of being more green and environmental-friendly. Moreover, experiments on relatively large-scaled problems allow us to demonstrate the potential benefits of the proposed heuristic     

面向对象的车辆路径问题仿真研究

针对车辆路径问题的复杂性和随机性,采用离散事件仿真的方法对车辆配送过程建模,并用面向对象技术实现了仿真模型。通过对象识别和层次划分,将模型分为实体建模和仿真控制建模两部分,构建了面向对象的车辆路径问题的仿真框架。这种方法克服了精确算法和智能启发式算法较难解决具有NP—hard性质的多约束车辆路径问题的弊端,实验结果验证了新方法的有效性和可行性。     

带时间窗约束的配载车辆调度问题研究

运用启发式算法求解带时间窗约束的配载车辆调度问题。借鉴多重旅行商问题的模型结构,建立配载车辆调度模型。通过改进C-K节约算法,在线路连接过程中插入时间窗约束和车辆容量约束的检验子程序,排除不符合约束条件的线路构形,构造求解调度模型的启发式算法。并给出算例,验证调度模型和启发式算法的可行性。     

A volume and material handling cost based heuristic for designing cellular manufacturing cells

One of the major problems in a group technology or cellular manufacturing environment is the formation of part groups and machine cells. Because of the combinatorial nature of the cell formation problem, it is difficult to solve the problem optimally. Most of the procedures related to cell design in cellular manufacturing operate on the part-machine incidence matrix in an attempt to identify block diagonality. If complete block diagonality does not exist, the decision about cell configuration is left to the subjective judgement of the designer. These procedures are also generally based on part routing only, and do not consider part volume and material handling costs.In this paper we develop an integer programming model, as well as a heuristic to effectively assign machines to cells. In these procedures we consider component volumes, costs related to movement of components between and within cells, and penalty for not using all machines in a cell visited by a component. Since the integer programming formulation becomes large even for small problems, an efficient heuristic is developed to solve larger problems. The heuristic solutions to 180 randomly generated small problems were compared against the optimal solutions obtained by the integer programming model. The heuristic has been found to identify optimal solutions in all 180 cases.This heuristic is also compared to several well known algorithms on 900 larger test problems. These problems were generated to cover a wide range of environmental situations such as varying levels of block diagonality in the part-machine incidence matrix, and diversity in the component volumes and material handling costs. In 99% of the problems our heuristic generated solutions which are better or as good as the best solution obtained by other algorithms. Further, in situations where complete block diagonality in the part-machine incidence matrix did not exist, our heuristic produced even better results. Since the maximum number of iterations required in our heuristic is the number of machines in the problem, the heuristic is computationally efficient.     

An evaluation of heuristic methods for determining the best table mix in full-service restaurants

Little research has been done on the optimal mix of supply in service businesses that maximizes revenue. Our research context is the full-service restaurant table mix problem. This problem, which is quite new to the literature, finds the optimal number of different size tables for a restaurant to maximize its value (revenue or contribution) generating potential. Specifically, we examine the effectiveness of eight heuristic techniques for the problem using two experiments. The first experiment uses data from a 240-seat full-service restaurant to evaluate all eight heuristics, while the second experiment investigates the performance of selected heuristics under a broader set of environmental factors. The results of our first experiment showed that the better of the simulated annealing heuristic variants yielded the optimal solution in seven of eight test problems, averaging within 0.1% of optimal. Our second experiment showed that the simplest of the models we investigated yielded solutions within 1% of the simulated annealing solution. Finally, we observed that altering the table mix on a daily basis increased performance by over 1% compared to maintaining the optimal weekly table mix.     

Interaction of maritime shipping and hinterland traffic using a two-level hierarchical transport network

A transport scenario based on a two-level network is developed in order to model operations in maritime transport. To enhance competitiveness and sustainability of shipping companies, the interaction between maritime shipping and hinterland traffic is considered in the novel two-level hierarchical transport network. Demand occurs on the second level, the hinterland, whereas the first level, the sea transport, acts as a pure distribution network. The two-level hierarchical transport model allows regulating the degree of interaction a liner operator has regarding the hinterland traffic. Moreover, access of vehicles to nodes or edges is given subject to feasibility constraints, relating to the size of a vehicle. In this work, the relation to existing routing models, such as the travelling salesman problem, the vehicle routing problem, or the travelling purchaser problem, is illustrated. Moreover, through the definition of a proper set of assumptions, it is possible to reduce the considered two-level hierarchical transport network to different standard routing models. These new insights allow the application of existing solution methods. Numerical experiments are conducted for a particular case and illustrate the influence of node costs on the distribution of workload to the vehicles.     

随机需求车辆路线问题研究

由现实生活中的不确定性因素引出需求随机的车辆路线问题,根据该问题随机性强的特点,设计了不断优化的求解方法。即首先用SFC和2-opt算法求解TSP以确定客户的访问顺序,并在车辆访问过程中不断对路线进行优化。最后,通过实例模拟验证该求解方法的可行性和适用性。     

VMI模式下的非周期性IRP问题

本文针对VMI管理模式下非周期性计划期、单车辆IRP问题，建立供应商库存管理路径优化模型．利用两阶段启发式算法结合蚁群算法、遗传算法等亚启发式算法来求解模型，并分析结果影响。     

基于真实路网的车辆路径问题研究

车辆路径问题(VRP)是广泛应用于物流配送的组合优化问题。在实际应用中,传统的处理方法是先利用路网信息求出任意两点间最短路,生成完全连通图,然后进行VRP的优化求解,这样会浪费时间和空间,降低求解效率。因此,提出了一种基于真实路网的VRP建模方法,把任意两点间最短路的求解与VRP联合起来,通过仿真实验与传统的处理方法进行了比较,两种方法得到了相同的最优解,但基于真实路网的VRP建模方法在计算时间方面占优。