基于自适应大邻域搜索算法的外卖配送车辆路径优化

目前我国外卖配送体量已达到单日数千万单的级别,外卖配送车辆的有效调度优化显得至关重要。针对外卖需求产生时间集中、配送时间要求严格的特点,设定顾客期望送餐时间窗以及时间惩罚成本,在考虑成本因素分析基础上,建立物流配送平台总成本最低为目标的数学模型。设计三种订单删除操作及两种订单插入操作,运用自适应大邻域搜索算法对不同规模算例进行求解。最后,利用CPLEX对算法结果进行最优验证,证明了算法的效率和精度。相关方法与结论可为即时配送调度优化提供决策支持。     

基于众包平台的外卖实时配送订单分配与路径优化研究

众包配送平台通过集结社会闲散运力,为应对激增的外卖实时配送需求提供了新的思路,其核心的订单分配与路径优化问题作为影响其配送成本与效率的关键问题受到关注。针对该问题中订单的实时性、时效性、配送员的自由性等特征,建立以平均每单配送距离以及平均每单完成时间最小为目标的实时订单分配与路径优化模型。分别设计了贪婪策略、最小差值策略用于求解该问题。最后通过大量的数值仿真研究验证了两个策略的有效性,发现最小差值策略所得的平均每单配送距离更短,贪婪策略所得的平均每单完成时间更短。进一步研究了两种策略在不同配送员容量限制、配送员数量、订单密度等参数变化时的适用性,需要控制成本宜采用最小差值策略,追求配送效率宜采取贪婪策略,研究结果可为众包配送平台的订单分配与路径优化策略的选择提供决策支持。     

基于感知等待时间的外卖O2O平台用户体验优化研究

目的 从时间维度中的用户体验和感知等待时间的视角出发,为降低服务等待过程中用户产生负面情绪的几率和完善现有O2O外卖类应用平台的用户体验提供新的优化方案.方法 基于交互设计和服务设计等相关理论及研究方法,以心理学中"感知等待时间"为理论分析切入点,首先结合影响感知等待时间的主客观因素来剖析O2O外卖服务中的等待时间要素,其次通过调研绘制出外卖O2O平台的用户体验地图,并重点挖掘在订单配送阶段中用户等待的痛点和机会点,最后结合需求层次论提出外卖O2O平台针对减轻感知等待时间的用户体验优化策略.结论 外卖O2O平台的用户体验并非仅为瞬时的界面交互感受,且感知等待时间是影响外卖O2O平台及其用户体验的重要因素,从时间维度重点优化订单配送阶段的用户体验可以缓解用户在等待过程中的不良心理感受,提升外卖O2O平台的整体服务.     

考虑顾客优先级的多目标O2O外卖即时配送路径优化研究

基于O2O外卖平台的配送现状,引入外卖平台顾客优先级概念,从顾客满意度和配送成本两个角度出发,建立了考虑客户优先级、带时间窗、动态、多车场多目标取送货车辆的路径模型,采用加权法将多目标转化成单目标,并设计改进的迭代局部搜索算法对模型进行求解,最后通过调研获取某校园周边外卖实际数据构造测试算例,通过数值实验,验证了模型和算法的有效性。实验结果表明:当成本权重超过一定数值时,成本下降空间变小,但顾客满意度会持续降低。     

基于众包的外卖配送订单选择研究

各外卖平台均提供了兼职配送员参与众包服务的渠道。与专职配送员相比,兼职配送员有着“路径开放、时间有限、最终目的地确定”等诸多不同的特点。基于兼职配送特点,为了提高众包配送员的接单效率,提高兼职收益,对众包模式下的订单选择及订单执行路径进行深入分析,提出将二者进行统一考虑的双层算法：在底层建立众包外卖配送路径规划模型,并使用改进的遗传算法求解;第2层利用贪心算法调用底层模型,通过比较配送收益进行订单选择,使得兼职人员的配送收益最大。通过算例实验,验证模型及算法的合理性及有效性条件,发现算法的计算时间随备选订单数量增加线性增加。在现实应用中,需要通过对备选订单进行打分排序,控制“订单池”规模,则能在可接受时间内得到较高质量的选择结果。     

考虑空载损失的非集中式共同配送订单分派及路径优化研究

物流活动的空载率居高不下源于路径规划不合理及企业间缺少合作,共同配送是降低空载损失的有效模式,但非集中式共同配送下物流企业可能基于被分派的订单选择自身成本最小的配送路径,从而导致共同配送联盟的空载损失变大。本文研究考虑空载损失的非集中式共同配送订单分派及路径优化,首先提出空载损失定义,权衡整个配送过程的成本最小和空载损失最小两个目标,基于非集中式共同配送的特征设计订单分派策略,进而建立订单分派及路径优化模型。设计了基于ε约束法的精确算法、改进的MOPSO (multiple objective particle swarm optimization)算法、多项式时间快速算法进行求解,并结合算例验证算法的有效性。数值分析结果表明,即使物流企业均追求自身成本最小化,提出的订单分派策略也可得到与全局优化相近的结果。     

质量体积双约束下车辆装载与配送路径联合优化研究

目的 针对质量与体积共同限制的配送路径问题,综合考虑订单不可拆分、货物的体积等约束,构建包含路径最短和装载率最高双目标的车辆装载与配送路径联合优化模型。方法 在车辆路径优化模型的求解方面,首先利用聚类算法对配送区域进行划分,然后通过车辆的载质量判断是否能进行站点货物的配送,最后利用遗传算法求得最优路径。在三维装载模型的求解上使用贪心算法和基于块的启发式算法,解决了货物的装箱问题。结果 基于某公司具体实例对模型与算法的可行性进行了验证,优化后配送的车辆减少了1辆,配送距离减少了154.247km,平均装载率达到了93.89%,节省了企业的配送成本。结论 所构建的模型以及求解的算法可以提高装载率和配送效率,为解决车辆装载与配送路径联合优化问题提供理论依据。     

基于动态需求的带时间窗的车辆路径问题

针对动态需求下的带时间窗的车辆路径问题,在最小化配送成本的目标下,通过提升服务的准时性来改进顾客满意度。考虑两阶段规划策略:在初始规划阶段,采用改进的遗传算法获得初始车辆路径;在动态优化阶段,将动态需求过程转化为多个瞬时静态子过程,采用模拟退火算法得到实时优化后的车辆路径方案。在一个实际案例中的应用和求解,证明了方法的现实有效性。     

车载可补货无人配送小车配送路径研究

无人配送小车由于不适合长距离运输,可与货车搭配完成“最后一公里”配送任务以增加服务范围,这对车辆路径优化问题提出了新的挑战。针对配送小车数量有限、城市配送货物量大且货车停靠限制的特点,提出无人配送小车可补货的大车-小车路径优化问题,即一辆货车搭载多台无人配送小车,由无人配送小车给客户送货,无人配送小车可在货车处补充货物并执行多行程配送。构建以总配送距离最短为目标的整数规划模型,针对此模型设计混合遗传大邻域搜索算法,在遗传算法基础上增加大邻域搜索算法对个体优化。在算法优化过程中先优化小车路径,再在小车路径基础上优化大车路径。数值实验表明,对于小规模问题,所提算法最多花费CPLEX求解时间的6%便获得最优解;在改造的Solomon数据上,所提算法相对于遗传算法平均有95.5%的计算结果优势,相对于大邻域搜索算法平均有7.2%的计算结果优势,且数据量越大,优势越大。     

基于改进贪婪关联算法的在线零售商优化拆单问题

针对在线零售商一地多仓及仅考虑品类拆单的场景,建立最大化整单配送模型,对单品分配方法进行研究,目的是通过改进现有算法优化配送中心中存放的单品,以进一步降低拆单率。针对贪婪订单算法和贪婪热销算法中未考虑单品间关系性的问题,结合Apriori算法,对算法进行优化设计,提出贪婪关联算法。算法应用一种新的单品分配方法寻求订单中具有强关联关系的单品,并对具有强关联关系的单品优先进行分配。实验结果表明,与贪婪订单算法和贪婪热销算法相比,改进后的算法能显著地降低拆单率,分别降低约8%和11%。     

带软时间窗的同时取送货车辆路径问题研究

考虑软时间窗下的车辆路径问题,客户点常伴有同时取送货的双重需求。针对此类问题,通过对软时间窗、车辆在途前后时间关系及二者融合问题进行刻画,同时将车辆行驶距离、车辆使用数、违反软时间窗总时间、客户满意度等纳入综合考量,构建相应混合整数非线性规划(mixed integer nonlinear programming, MINLP)模型。设计相应多目标优化求解算法,运用理想点法对目标函数进行转化,将多目标优化问题转化为单目标优化问题。结合相应算例集,运用LINGO 17.0全局求解程序求得每组算例的全局最优解。结果表明,针对带软时间窗的同时取送货车辆路径问题(vehicle routing problem with simultaneous pick-up and delivery and soft time windows, VRPSPDSTW),所建模型及算法是有效且可行的。     

Risk approaches for delivering disaster relief supplies

We consider the problem of designing the logistic system to assure adequate distribution of relief aid in a post-natural-disaster situation, when damages to infrastructure may disrupt the delivery of relief aid. The problem is formulated as a multi-objective optimization problem, encompassing three objective functions of central interest in such problems. The first objective function is a measure of risk (various forms of such risk are analyzed). The second objective function measures the coverage provided by the logistic system in the distribution of relief aid to disaster victims. The third objective function represents total travel time. We focus on the risk of delivery tours for relief supplies, where risk here captures the threat that potential tours become impassable after the natural hazard event. In order to cope with a range of different natural disasters and policy objectives, we develop five approaches emphasizing different measures of tour-dependent risk. To cover both earthquake and flood risks, we consider correlated as well as uncorrelated risk measures. We develop a two-phase solution approach to reflect the dictates of real-world disaster relief motivating this analysis. The first phase generates potentially Pareto-optimal solutions to the overall multi-objective logistic design problem with respect to three objectives. For any given risk measure, the first-phase design problem is formulated as a multi-objective integer program and a memetic algorithm is proposed as the solution approach. The second phase is an enrichment procedure to generate a broader range of potentially Pareto-optimal alternatives. The suggested approach is tested on real-world data from the province of Manabí in Ecuador and the results associated with the different risk measures are analyzed to illustrate the value of the proposed approach for the design of disaster relief networks.     

Multi-objective optimization of engineering systems using game theory and particle swarm optimization

This study proposes particle swarm optimization (PSO) based algorithms to solve multi-objective engineering optimization problems involving continuous, discrete and/or mixed design variables. The original PSO algorithm is modified to include dynamic maximum velocity function and bounce method to enhance the computational efficiency and solution accuracy. The algorithm uses a closest discrete approach (CDA) to solve optimization problems with discrete design variables. A modified game theory (MGT) approach, coupled with the modified PSO, is used to solve multi-objective optimization problems. A dynamic penalty function is used to handle constraints in the optimization problem. The methodologies proposed are illustrated by several engineering applications and the results obtained are compared with those reported in the literature.     

动态视角下城市通勤路径优化研究

针对城市上班族“通勤难”的问题,考虑到交通拥堵发生的不确定性,以动态视角研究基于实时交通信息的通勤车动态路径优化问题,同时将停靠点选址问题作为其影响因素优先进行分析。建立以行驶路径最短为目标的初始路径优化模型,并重新制定路线更新规则,将路口交叉点作为新的关键点引入更新策略。通过实证分析得到,动态优化后无拥堵状态下的通勤效率可以提高35.7%,存在拥堵的状态下通勤效率可以提高40%,证明了模型和算法的有效性。     

Balancing multiple criteria in formulation of weighted,single-objective genetic algorithm optimization for CNC machining problems

This paper presents results obtained from the implementation of a genetic algorithm (GA) to a simplified multi-objective machining optimization problem. The major goal is to examine the effect of crucial machining parameters imparted to computer numerical control machining operations when properly balanced conflicting criteria referring to part quality and process productivity are treated as a single optimization objective. Thus the different combinations of weight coefficient values were examined in terms of their significance to the problem's response. Under this concept, a genetic algorithm was applied to optimize the process parameters exist in typical; commercially available CAM systems with significantly low computation cost. The algorithm handles the simplified linear weighted criteria expression as its objective function. It was found that optimization results vary noticeably under the influence of different weighing coefficients. Thus, the obtained optima differentiate, since balancing values strongly affect optimization objective functions.     

Effective multi-objective optimization with the coral reefs optimization algorithm

In this article a new algorithm for multi-objective optimization is presented, the Multi-Objective Coral Reefs Optimization (MO-CRO) algorithm. The algorithm is based on the simulation of processes in coral reefs, such as corals' reproduction and fight for space in the reef. The adaptation to multi-objective problems is a process based on domination or non-domination during the process of fight for space in the reef. The final MO-CRO is an easily-implemented and fast algorithm, simple and robust, since it is able to keep diversity in the population of corals (solutions) in a natural way. The experimental evaluation of this new approach for multi-objective optimization problems is carried out on different multi-objective benchmark problems, where the MO-CRO has shown excellent performance in cases with limited computational resources, and in a real-world problem of wind speed prediction, where the MO-CRO algorithm is used to find the best set of features to predict the wind speed, taking into account two objective functions related to the performance of the prediction and the computation time of the regressor.     

A meshless Pareto-optimal method for topology optimization

A meshless Galerkin Pareto-optimal method is proposed for topology optimization of continuum structures in this paper. The compactly supported radial basis function (CSRBF) is used to create shape functions. The shape function is constructed by meshfree approximations based on a set of unstructured field nodes. Considering the Pareto-optimality theory, the initial single objective topology optimization problem is transformed into multi-objective problem. The optimum solution is traced via the Pareto-optimal frontier in a computationally effective manner. The optimal problem does not need to be solved directly. Finally, several examples are used to prove the validity and effectiveness of the proposed approach.     

A study on evolutionary multi-objective optimization for flow geometry design

In this paper, we investigate three recently proposed multi-objective optimization algorithms with respect to their application to a design-optimization task in fluid dynamics. The usual approach to render optimization problems is to accumulate multiple objectives into one objective by a linear combination and optimize the resulting single-objective problem. This has severe drawbacks such that full information about design alternatives will not become visible. The multi-objective optimization algorithms NSGA-II, SPEA2 and Femo are successfully applied to a demanding shape optimizing problem in fluid dynamics. The algorithm performance will be compared on the basis of the results obtained.     

A new multi-objective particle swarm optimizer using empirical movement and diversified search strategies

Most real-world optimization problems involve the optimization task of more than a single objective function and, therefore, require a great amount of computational effort as the solution procedure is designed to anchor multiple compromised optimal solutions. Abundant multi-objective evolutionary algorithms (MOEAs) for multi-objective optimization have appeared in the literature over the past two decades. In this article, a new proposal by means of particle swarm optimization is addressed for solving multi-objective optimization problems. The proposed algorithm is constructed based on the concept of Pareto dominance, taking both the diversified search and empirical movement strategies into account. The proposed particle swarm MOEA with these two strategies is thus dubbed the empirical-movement diversified-search multi-objective particle swarm optimizer (EMDS-MOPSO). Its performance is assessed in terms of a suite of standard benchmark functions taken from the literature and compared to other four state-of-the-art MOEAs. The computational results demonstrate that the proposed algorithm shows great promise in solving multi-objective optimization problems.     

数字化车间物料动态配送规划研究

目的 针对数字化生产车间工位物料需求时间的不确定,导致物料配送不准确、不及时的问题,提出一种动态物料配送策略。方法 首先,根据工位关联度和变动时间窗确定实时的配送工位和协同配送工位,设计基于工位排序的动态物料配送路径优化策略。其次,建立以配送成本和时间窗偏离惩罚成本综合最小为目标函数的数学模型。最后,提出并采用系统动力学仿真与蚁群遗传融合算法联合的方法对模型进行求解。结果 模拟算例表明,与静态物料配送优化策略相比,该策略的平均时间成本减少率为30.1%,平均库存减少率为14.86%。结论 该策略能够根据动态时间窗确定配送工位和协同工位,并实时调整配送顺序,实现物料配送的动态自适应性调整,降低总配送成本。融合算法在迭代次数、收敛性、最优解质量方面有明显优越性。