基于K-means聚类算法与重心法的故障共享单车回收中心选址优化

共享单车已日渐成为人们短途出行的重要交通工具，但共享单车市场一贯采用“重投放，轻维护”的发展模式，而共享单车使用中的正常损耗及人为破坏等造成的故障车数量却十分庞大，对其回收修复或报废的任务繁重，这也成了共享单车市场逆向物流亟待解决的难题。该文立足武汉共享单车市场，通过对故障共享单车报废点的聚类分析，基于运输成本导向，使用重心法探寻回收中心最佳选址点，以解决故障共享单车回收成本和效率问题。对模型的模拟验证表明，优化后的回收中心选址点不仅能降低故障共享单车回收成本，而且相比武汉市现有的三个分布较远的回收中心，其总体运营成本更低，故障共享单车回收效率更高，便于共享单车的分区域运营管理。事实证明，基于K-means聚类算法与重心法确定回收中心选址问题不仅操作简单，可行性高，而且方便快捷，相较于现实中单一考虑成本等的选址方式，此模型更能兼顾多方面因素，优势明显。基于K-means聚类算法与重心法来确定回收中心选址，适用于城市的各个区域，选点精确又方便高效，模型具有较强的实用性。     

共享单车调度优化问题研究

针对共享单车站点经常出现的供需不平衡问题,提出人工调度策略,以提高单车利用率和用户满足率.首先将一天划分为几个用车高峰时段,根据每个站点的单车使用历史数据,计算各站点在每个时段的需求量区间;在区域内单车总投放量不变的前提下,基于每个时段初期各个站点存放的单车数量,确定单车调出站点和单车调入站点,进一步以站点之间的单车调度数量为决策变量,建立共享单车调度问题的整数规划模型,使区域内各个单车站点的供需量基本达到平衡,并且总调度成本最小.利用北京市海淀区共享单车数据进行模拟计算,对比分析了调度优化前后的共享单车利用率和用户满足率.结果显示,调度优化后,单车利用率平均提高7.78%,用户满足率平均提高13.09%;综合考虑企业调度成本和收入情况可以发现,通过调度优化,企业的平均利润增长率为7.53%.本文的研究结果可以帮助共享单车企业提升管理水平,增加利润.     

中转点变化的应急医疗物资联合运送干扰管理研究

应急响应中常常出现供应节点与需求节点距离太远、关键道路损毁等情境,此时直升机和车辆逐渐被联合使用以运送关键应急物资。由于原生及次生衍生灾害的发生发展,可能导致已有应急中转点不能使用,或者由于救援工作的开展,产生新的应急中转点等,这就可能导致事先已经制定的联合运送方案不能按照原计划进行。针对直升机和车辆联合运送中出现的中转点变化干扰事件,从物资到达时间、联合运送路线和使用运力三方面进行扰动度量,进而建立一个中转点变化的应急医疗物资联合运送干扰恢复模型,并通过改进基于客户的编码方法和提出基于简化策略的种群初始化方法等,设计了模型求解的遗传算法。数值实验验证了模型与算法的有效性,并通过与重优化方法对比,检验了干扰管理方法在处理中转点变化干扰事件中的优势。     

考虑多仓库的共享单车重新配置问题研究

共享单车是我国大力提倡的低碳交通出行模式,加快共享单车发展是解决最后一公里、城市拥堵和环境污染等问题的重要途径。由于人们停放共享单车的无规律性,使得共享单车系统中各车桩的单车库存量存在不平衡。如何合理的对车桩中的单车进行重新调配,来满足用户的需求,是相关企业亟待解决的问题。共享单车的调配路线优化是优化车桩库存量的重要手段之一。本文研究多仓库条件下的货车调配路线优化问题,建立了一个混合整数非线性规划模型。不同于传统的路径优化问题的研究大多是以成本或时间为目标,本文采用基于车桩库存量的非线性惩罚函数来表示用户需求,从而使得所研究的问题是一个凸函数优化问题。为了简化本文的问题,将目标函数分段线性化。基于车桩网络的特点,设计了变邻域搜索算法,以及构建初始解的贪婪算法。最后,以某共享单车公司为例,进行算例分析,来说明模型和算法的合理性和有效性。     

不确定需求下故障共享单车回收周期性车辆路径问题研究

为了及时有效地回收城市道路网络中的故障共享单车,对分散于路网边上的故障单车进行聚类形成收集点,考虑聚类收集点上回收需求呈现的不确定性特征,建立以行驶总距离最小为目标的回收周期性车辆路径选择模型.采用基约束鲁棒优化方法,利用有界区间对不确定的回收量进行描述,引入扰动系数和控制系数调节模型的鲁棒性和适应性.针对模型设计近似...     

最小化碳排放的共享单车迁移问题

本文考虑共享单车迁移问题, 它可看作是经典旅行售货商问题的一个新颖变形, 不同的是其目标函数为最小化碳排放。其中, 碳排放利用单车负载与其行驶路程的乘积进行刻画。我们提出了两个启发式算法：贪心和基于TSP的算法, 每个算法的核心思想均是优先减少单车负载。从理论上证明算法的可行性并给出数据实验以验证算法的实际性能。数据实验结果表明贪心算法优于基于TSP的算法, 这为共享单车企业进行日常单车分配提供了理论依据。     

最小化碳排放的共享单车迁移问题

本文考虑共享单车迁移问题, 它可看作是经典旅行售货商问题的一个新颖变形, 不同的是其目标函数为最小化碳排放。其中, 碳排放利用单车负载与其行驶路程的乘积进行刻画。我们提出了两个启发式算法：贪心和基于TSP的算法, 每个算法的核心思想均是优先减少单车负载。从理论上证明算法的可行性并给出数据实验以验证算法的实际性能。数据实验结果表明贪心算法优于基于TSP的算法, 这为共享单车企业进行日常单车分配提供了理论依据。     

政府激励、质量投入对共享单车服务网络均衡的影响研究

本文在考虑服务供应链闭环网络系统中同类成员竞争的基础上，将共享单车服务网络划分为制造供应商、服务集成商和消费需求端三个层次，建立考虑政府激励、质量偏好的共享单车服务网络均衡模型。结合修正投影收缩法，设计模型求解程序进行数值算例分析。研究发现：服务集成商新产品价格随着平均共享频率变化而变化，再制造产品价格不仅受平均共享频次的影响，还与政府给予的激励存在关联，当政府给予激励越大，其内生价格越低；服务集成商回收产品内生价格与服务集成商的质量偏好存在负向关联关系，而与制造供应商的质量偏好存在正向关联关系；共享频次是共享服务闭环供应链实现可持续发展的关键指标。     

面向智能合约的共享单车供应链产品质量与投放决策研究

近年来，共享单车行业不断扩张，形成了以互联网平台企业为主导的寡头垄断市场，当前共享单车供应链中存在着供应商话语权缺失、消费者信息易泄露、平台企业资金压力大等问题。基于此，本文构建了面向区块链智能合约技术的新型共享单车供应链，将其与现有基于互联网平台的共享单车供应链在单车投放量、产品质量以及供应链收益水平等方面进行比较研究，分析当前供应链企业做出模式创新的市场环境条件，研究不同市场因素对不同供应链的影响。通过对共享单车市场竞争的仿真模拟，验证了博弈模型的分析结论。研究表明，智能合约技术的引入能够减少共享单车上下游的过度竞争，同时提高行业服务质量及便利性。     

通行受限下需求可拆分的应急物资卡车-多无人机协同配送路径优化

研究通行受限情景下需求可拆分的应急物资卡车-多无人机协同配送路径优化问题,综合考虑灾区路网状况、卡车可途中发射/接受无人机、无人机单次起飞可配送多个需求点、需求可拆分等因素,以应急物资配送任务完成时间最短为目标,构建卡车-多无人机协同配送路径优化模型.根据问题与模型特征设计一种改进蚁群算法求解.实验结果表明:文章方法能合理分配卡车与无人机的配送任务,科学规划通行受限情景下需求可拆分的应急物资卡车-多无人机协同配送路径;卡车途中发射/接收无人机方式能有效缩短无人机飞行距离,减少卡车与无人机的协同时间,缩短通行受限情景下的应急物资配送时间,具有可行性、合理性与有效性.     

多类型公共自行车调运问题

公共自行车是我国正大力发展的低碳交通出行模式,加强公共自行车调运优化是提升自行车出行吸引力的关键要素。通过对公共自行车调运背景分析,提出了一类多类型公共自行车的调运优化问题。针对现实生活中租赁站点内公共自行车不均衡的情况,建立了以总成本最小为目标的混合整数线性规划模型,并提出一种改进的混合禁忌搜索对问题进行求解。通过数值实验分析了问题特性并验证了算法性能。实验结果表明非均衡惩罚系数决定了租赁站点各类自行车的装卸载数量,并影响了调配车辆的运行路线,是实现多类型公共自行车均衡优化的关键因素。不同类型自行车的替代策略使得调运决策更加灵活。混合禁忌搜索可以求解更大规模的问题,并能在短时间内求得较好质量的解。     

A two-phase heuristic approach to the bike repositioning problem

An approach to overcome the bike imbalance problem is to transfer excess bikes to branches with bike shortages. This study develops a constrained mathematical model to deal with a multi-vehicle bike-repositioning problem, and aims to minimize the sum of transportation and unmet demand costs over a planning horizon through bike-transfer strategies under a minimum service requirement. A two-phase heuristic based on linear programming was proposed to solve the problem and produce compromising solutions. In the first phase, the paper developed a linear programming model to quickly develop decisions related to bike inventory, unloading, and loading for all stations for each time slot. In the second phase, this paper proposed an iterative approach through two parameter sensitive mathematical models to sequentially reduce the problem scale to develop decisions related to bike transfers. Computational results show that the proposed approach is superior to a CPLEX optimizer and a hybrid heuristic based on a genetic algorithm. The proposed approach was used to analyze the bicycle system in Taiwan. The impacts of various system parameters on the system were also investigated.     

A multi-periodic optimization formulation for bike planning and bike utilization

The number of policy initiatives to promote the use of bike, or the combined use of bicycle and public transport for one trip, has grown considerably over the past decade as part of the search for more sustainable transport solutions. This paper presents an optimization formulation to design a bike-sharing system for travel inside small communities, or as a means to extend public transport for access and egress trips. The mathematical model attempts to optimize a bike-sharing system by determining the minimum required bike fleet size that minimizes simultaneously unmet demand, unutilized bikes, and the need to transport empty bikes between rental stations to meet demand. The proposed approach is applied to an example problem and is shown to be successful, ultimately providing a new managerial tool for planning and analyzing bike utilization more effectively.     

The static bike relocation problem with multiple vehicles and visits

This paper introduces the static bike relocation problem with multiple vehicles and visits, the objective of which is to rebalance at minimum cost the stations of a bike sharing system using a fleet of vehicles. The vehicles have identical capacities and service time limits, and are allowed to visit the stations multiple times. We present an integer programming formulation, implemented under a branch-and-cut scheme, in addition to an iterated local search metaheuristic that employs efficient move evaluation procedures. Results of computational experiments on instances ranging from 10 to 200 vertices are provided and analyzed. We also examine the impact of the vehicle capacity and of the number of visits and vehicles on the performance of the proposed algorithms.     

Adapting the ng-path relaxation for bike balancing problems

This paper deals with a static bike relocation problem that deploys a fleet of vehicles to redistribute shared bicycles. To solve the problem to optimality, we present a branch-price-and-cut algorithm. In particular, a new path relaxation for the pricing problem is introduced that relaxes the constraints on the maximum number of bikes to move at each station in a similar fashion as elementary can be relaxed in vehicle routing problems. Computational results show that our algorithm outperforms the former state-of-the-art.     

Scheduling trucks in local depots for door-to-door delivery services

A scheduling method is suggested for trucks delivering and picking up freight between branch offices and a regional depot in door-to-door delivery services. As the objective functions, different levels of customer service resulting from different timing of deliveries and pickups to/from branch offices are considered as well as the travel cost of trucks. Useful properties of the optimal timing of deliveries and pickups are derived to reduce the size of the search space significantly. Numerical experiments were conducted to evaluate various algorithms to solve the problem.     

�����������Ĵ������г�����ϵͳ����վ�㳵����ƽ���о�

??Recently, big data, could computing and internet of things provide some new information technologies for organization and management of complex systems, and they have caused multifaceted changes on organization framework and operations mechanism of enterprises. Based on this, we first construct a new stochastic model for a big data driven large-scale bike-sharing system, which expresses the important role played by big data, and describes the operations mechanism of the large-scale bike-sharing system, and specifically, the rebalancing of bikes in various stations in terms of trucks. Then, we present a mean-field limit theory, which is applied to analyzing the big data driven large-scale bike-sharing system, including establishing a time-inhomogeneous queueing system by means of the mean field theory, and setting up the mean-field equations through the time-inhomogeneous queueing system; providing an empirical measure process by means of a nonlinear birth-death process, giving algorithms for computing the fixed point in terms of a segmented structural birth-death processes, and computing the average number of bikes in each station; and providing numerical examples to analyze how the steady average number of bikes in each station depends on some key parameters of the bike-sharing system. Using these results, this paper analyzes physical effect of big data on performance of the large-scale bike-sharing. Therefore, this paper gives a promising research direction of stochastic model in the study of large-scale bike-sharing systems.     

集装箱码头堆场翻箱与外集卡提箱顺序同步优化方法

针对集装箱码头提箱作业过程中,由于外集卡的提箱顺序与目标箱在堆场的堆存位置不匹配导致大量翻箱这一难题,以码头的作业成本和外集卡的延误成本之和最小为目标,建立堆场翻箱与外集卡提箱顺序同步优化模型,优化外集卡的提箱顺序、龙门吊的任务分配以及翻箱方案。设计基于动态规划的启发式算法求解模型,并利用算例对模型与算法的有效性进行了验证。结果表明:与目前码头普遍采用的提箱方式相比,通过调整外集卡提箱顺序并同时优化翻箱方案以及龙门吊的任务分配可以降低堆场翻箱率,减少龙门吊的移动成本,从而节省提箱作业的总成本。