Traffic Speed Interface Development in Route Choice Decision

In this paper,a method has been developed based on historic traffic data(speed),which helps the commuters to choose routes by their intelligence knowing the traffic conditions in Google maps.Data has been collected on basis of video analysis from several segments between Tuker Bazar and Bandar Bazar route.For each of the video footage,a reference length has been recorded with measurement tape for use in video analysis.A software has been also developed based on Java language to get the traffic information from historic data,which shows the output as images consisting of traffic speed details on the available routes by giving day and time limit as inputs.The developed models provide useful insights and helpful for the policy makers that can lead to the reduction of traffic congestion and increase the scope of intelligence of the road users,at least for the underdeveloped or developing country where navigation is still unavailable.

Stochastic Route Choice Equilibrium Assignment for Travelers with Heterogeneous Regret Aversions

Most route choice models assume that people are completely rational.Recently,regret theory has attracted researchers'attentions because of its power to depict real travel behavior.This paper proposes a multiclass stochastic user equilibrium assignment model by using regret theory.All users are differentiated by their own regret aversion.The route travel disutility for users of each class is defined as a linear combination of the travel time and anticipated regret.The proposed model is formulated as a variational inequality problem and solved by using the self-regulated averaging method.The numerical results show that users'regret aversion indeed influences their route choice behavior and that users with high regret aversion are more inclined to change route choice when the traffic congestion degree varies.

Fear Factor: Level of Traffic Stress and GPS Assessed Cycling Routes

Background: Cycling currently comprises only 1% of transport trips in the U.S. despite benefits for air pollution, traffic congestion, and improved public health. Methods: Building upon the Level of Traffic Stress (LTS) methodology, we assessed GPS trip data from utilitarian cyclists to understand route preferences and the level of low stress cycling connection between origins and destinations. GPS data were obtained from adult transport cyclists over multiple days. All bikeable road segments in the network were assigned an LTS score. The shortest paths between each origin and destination along bikeable roadways and along low stress (LTS 1 or 2) routes were calculated. Route trajectories were mapped to the LTS network, and the LTS and distances of observed, the shortest and low stress routes were compared. LTS maps and animations were developed to highlight where low stress connections were lacking. Results: There were 1038 unique cycling trips from 87 participants included in the analysis. An exclusively low stress route did not exist for 51% of trips. Low stress routes that were possible were, on average, 74% longer than the shortest possible path and 56% longer than the observed route. Observed routes were longer and lower stress than the shortest possible route. Conclusions: Results indicate that transport cyclists traveled beyond low stress residential areas and that low stress routes with acceptable detour distances were lacking. Cyclists appeared to weigh both route distance and quality and were willing to trade maximum directness for lower stress. GPS data provide additional information to support planning decisions to increase the impact of infrastructure investments on cycling mode share.

Particle swarm optimization for train schedule of Y-type train routing in urban rail transit

The train schedule usually includes train stop schedule,routing scheme and formation scheme.It is the basis of subway transportation.Combining the practical experience of transport organizations and the principle of the best match between transport capacity and passenger flow demand,taking the minimum value of passenger travel costs and corporation operating costs as the goal,considering the constraints of the maximum rail capacity,the minimum departure frequency and the maximum available electric multiple unit,an optimization model for city subway Y-type operation mode is constructed to determine the operation section of mainline as well as branch line and the train frequency of the Y-type operation mode.The particle swarm optimization(PSO)algorithm based on classification learning is used to solve the model,and the effectiveness of the model and algorithm is verified by a practical case.The results show that the length of branch line in Y-type operation affects the cost of waiting time of passengers significantly.

多出行场景下城市路内停车影响因素分析与行为建模

城市路内停车行为选择受到费用、距离、停车供给、道路状况等多种因素影响,不同出行场景下各种因素的影响程度也存在明显差异。为挖掘多种场景下影响居民路内停车行为因素之间的差异,以指导城市停车规划和管理政策的制定,首先,基于SP调查数据,对出行场景进行划分,并融合多源时空停车大数据,对不同场景下的路内停车行为影响因子进行提取;其次,使用离散选择模型对停车行为进行模拟,分析不同场景下各个因子对路内停车行为的影响程度。研究结果显示:停车距离对各类场景下的路内停车行为都具有显著影响;停车时长对商务和休闲场景下的路内停车行为影响较大;在通学、就医等场景下,各因素对路内停车行为的影响都不显著。

A DYNAMIC MODEL OF MULTIPLE PATH TRAFFIC ASSIGNMENT

A “Random Shortest Path”traffic assignment model and its algorithm arepresented by simulating the trip-makers’route-choice characters,and the dynamic meth-od is introduced in the assignment model.It is a ideal multiple path assignment modelwhich can be carried out by the dynamic method and static method,can better reflect boththe shortest path factor and the random factor in the route-choice,and is of reasonableassignment volumes.Besides,both dynamic and static softwares particularly suited to thetraffic assignment of large and medium-sized transportation networks arc developed.

基于机器学习的中型城市居民出行方式选择行为研究

为探索中型城市居民出行特征以及不同因素对出行方式选择行为的影响机制,本文以中国某中型城市居民出行数据为例,综合考虑传统离散选择模型和机器学习模型在预测精度和建模合理性上的优劣,以及机器学习模型超参数求解算法的特点和效率,引入变异程序,提出粒子群优化随机森林的中型城市居民出行方式选择预测模型,采用预测准确率、出行方式比例预测绝对误差和期望模拟误差这3项性能指标,量化对比粒子群优化随机森林模型与多种机器学习模型和多项Logit模型统计学上的预测性能差异,利用SHAP(SHapley Additive exPlanation)模型深入分析个人社会经济属性、出行属性及出行方式属性等相关因素与居民出行方式选择之间的非线性关系。结果表明:粒子群优化随机森林模型整体平均预测准确率最高,为0.856,出行方式比例预测平均绝对误差和期望模拟平均误差最低,分别为0.062和0.306,模型间指标差异在统计学检验下显著;距离对不同出行方式选择的影响最显著,步行和私家车出行对距离敏感性更高,不同距离下,两者选择概率变化超过35%;30岁以下群体不同出行方式选择概率差距大于其他年龄段;性别、是否拥有私家车或公交IC卡等因素显著改变公交车和私家车的选择概率。

不确定环境下收费与信息对路径选择行为影响的实验研究

在不确定环境下考虑信息发布与拥挤收费两种措施,基于交通网络均衡,推导了两路径简单路网的最优收费模型,并推广至一般路网。基于两路径路网设计逐日路径选择行为实验,分析了两种措施对路径选择行为的影响。结果表明:完美信息与收费的实施均能减少流量波动,单独实施收费的效果最佳,但只有两种措施相结合时,路网流量趋于用户均衡的趋势最稳定;完美信息会增加路径切换行为,而收费可有效抑制大幅的路径切换;无收费时路径切换往往会造成出行成本增加,而收费时则相反;收费与信息结合时的出行时间比两种措施单独实施时小。

外卖配送电动自行车换电柜选址定容实证研究

针对外卖配送电动自行车换电柜布局不合理带来的部分换电柜利用率低与部分换电需求得不到及时满足的供需矛盾问题,本文通过聚类POI(Point of Interest)数据确定外卖配送起止点,并通过仿真模拟外卖骑手配送路径预测外卖配送电动自行车换电需求时空分布,构建换电柜运营商总成本最低和用户满意度最高的多目标换电柜选址定容模型,并以新乡市主城区为例,采用NSGA-II(Non-dominated Sorting Genetic Algorithm II)算法得到换电柜选址定容方案。研究结果表明:仿真模拟得出的换电需求时间分布预测值与实际值基本吻合,换电需求在11:00,14:00,17:00和20:00左右急剧增长,且11:00和14:00左右的换电需求量显著高于17:00和20:00左右的换电需求量,外卖骑手配送路径仿真模拟方法在换电需求预测上具有较高的预测精度;换电柜选址方案不能同时满足运营商和用户利益均为最优,用户满意度的提高需以增加运营商总成本为代价;同时,兼顾运营商和用户利益的新乡市主城区外卖配送电动自行车换电柜最佳建设数量为26,其中,容量为11的换电柜11个,容量为22的换电柜8个,容量为33的换电柜7个;新乡市主城区应按照备选点编号15-7-19-17依次新增换电柜至30个,此时,用户满意度最大,若继续增加换电柜建设数量,只会增加运营商总成本。

考虑暴雨灾害动态影响的城市应急车辆救援路径优化研究

近年来,极端天气事件发生频次不断增加,强度不断加大,其中,由暴雨引发的城市内涝导致交通应急事件发生概率进一步增大。为提升暴雨灾害下应急救援响应速度,本文开展应急车辆救援路径优化研究。以通行时间最短为目标,考虑路面积水对车辆通行速度的动态影响,构建应急车辆救援路径优化模型,提出动态最短路径优化算法求解模型。选取上海市长宁区东北部作为研究区域,根据SWMM(Storm Water Management Model)模拟得到的50年一遇暴雨条件下城市道路路面的积水情况,设定应急救援场景,求解应急救援路径。通过本文提出算法求解得到的路径与传统静态最短路径算法求解结果对比可知,通行用时同比减少了25.42%。同时,考虑应急物资储备情况分配应急救援任务,扩展了算法的应用场景,形成可靠和高效的应急响应方案,可为提升暴雨灾害下应急响应效率提供参考。

Framework,model and algorithm for the global control of urban automated driving traffic

Automated driving has recently attracted significant attention.While considerable research has been conducted on the technologies and societal acceptance of autonomous vehicles,investigations into the control and scheduling of urban automated driving traffic are still nascent.As automated driving gains traction,urban traffic control logic is poised for substantial transformation.Presently,both manual and automated driving predominantly operate under a local decision-making traffic mode,where driving decisions are based on the vehicle’s status and immediate environment.This mode,however,does not fully exploit the potential benefits of automated driving,particularly in optimizing road network resources and traffic efficiency.In response to the increasing adoption of automated driving,it is essential for traffic bureaus to initiate proactive dialogs regarding urban traffic control from a global perspective.This paper introduces a novel global control mode for urban automated driving traffic.Its core concept involves the central scheduling of all autonomous vehicles within the road network through vehicle-infrastructure cooperation,thereby optimizing traffic flow.This paper elucidates the mechanism and process of the global control mode.Given the operational complexity of expansive road networks,the paper suggests segmenting these networks into multiple manageable regions.This mode is conceptualized as an autonomous vehicle global scheduling problem,for which a mathematical model is formulated and a modified A-star algorithm is developed.The experimental findings reveal that(i)the algorithm consistently delivers high-quality solutions promptly and(ii)the global scheduling mode significantly reduces traffic congestion and equitably distributes resources.In conclusion,this paper presents a viable and efficacious new control mode that could substantially enhance urban automated traffic efficiency.

考虑时空协同优先的城市应急车辆路径优化方法

目前城市应急救援车辆救援效能提升面临交通拥堵、交叉口延误、突发事故等诸多挑战。传统路径算法忽略了实际救援工作需求和路径选择偏好特征,无法很好满足快速高效、安全可靠的救援需求。针对这一问题,立足应急车辆与社会车辆在优先通行政策、车辆特性、优化目标、路径选择偏好等多特征差异,提出一种考虑时空协同优先的两阶段应急车辆救援路径动态优化方法。第1阶段,将用于求解静态K最短路径的Yen算法与城市路网时空动态性、高度随机性及不确定性等特征结合,增加基于层次聚类算法的路径安全风险约束与基于偏斜度指数等综合评价指标的可靠性约束,构建带有安全可靠性约束的动态K最短路径精确算法。第2阶段,结合车道数量、路段限速、路段长度等静态属性特征与拥堵延迟指数、路段平均车速等动态属性特征,求解在时空协同优先条件下的最佳救援路径。结果表明:与历史救援轨迹相比,考虑时空协同优先的两阶段城市应急车辆路径优化方法能够有效缩短城市应急车辆的救援响应时间,并提高应急车辆救援安全性与可靠性。因此,本方法面向实际救援工作,具有较好的应用前景。

考虑心理潜变量的城市轨道站点接驳方式选择行为研究

本文以计划行为理论为基础框架,从出行者使用接驳工具的经验和对接驳工具属性评价两个维度分别引入出行习惯、感知态度两类潜变量,构建城市轨道站点接驳方式选择意向的多指标多因果模型,探究城市轨道站点接驳方式选择意向的影响因素及其作用路径,并在此基础上结合巢式Logit模型构建混合选择模型,探究出发端与到达端接驳方式选择行为的影响因素及其影响机理。结果表明:行为习惯潜变量直接影响接驳方式选择意向,感知态度对步行、共享单车接驳选择意向具有直接影响;显著影响接驳方式选择意向的可观测变量有学历、车辆保有情况、收入情况、接驳距离及出行人数;相较于慢行交通与道路公共交通,私人小汽车、电动车的有无对私人交通方式接驳选择行为影响较大;学历和职业等个人社会经济属性,出行目的和出行人数等出行属性,以及主观规范和行为意向等心理属性对出行者城市轨道出发端和到达端接驳方式选择行为均有显著影响;相较于到达端,出行费用、主观规范及行为态度对出发端接驳方式选择行为的影响更大,且出发端接驳方式选择对到达端的接驳决策行为具有影响。

考虑充电策略的冷链物流配送路径优化研究

随着绿色与可持续发展要求的提出,使用电动物流车进行冷链物流配送逐渐成为热点,在考虑充电策略、车辆载重和客户时间窗等约束下,构建了配送总成本最小化为目标的电动车辆在冷链物流配送中的路径优化模型。根据设计的模型特点,提出了一种结合海洋捕食者和蚁群算法的混合算法进行求解,该算法有效提高了搜索能力和全局信息捕捉。根据算例分析对比可知,与完全充电策略相比,采用部分充电策略可以提高配送效率,有效降低了17.34%物流成本。通过设置车辆最大载重和不同充电站排队时间,分析车辆最大载重和充电站排队时间对物流总成本的影响,从而为企业提供不同车辆的选择。用实际案例和具体数据进行实验,验证了构建模型的有效性,证明了算法的高效性。

多车场多车型冷链物流时变绿色车辆路径优化

为降低运输成本,减少环境污染,提出一种考虑速度时变的多车场多车型绿色车辆路径优化方法。针对城市交通拥堵路况常发的现状,引入百度智慧交通网预测的拥堵指数反映实时路况,兼顾经济成本和环境成本,构建以总运输成本最小为优化目标的时变绿色车辆路径模型。根据模型特点改进遗传算法,并运用大中小3种规模的算例对其进行仿真实验,验证了模型和算法可以有效地降低物流配送成本,促进配送车辆节能减排,为生鲜电商冷链物流配送决策提供车辆运输路径方案。

城市轨道交通网络跨线列车开行方案优化模型研究

通过开行跨线列车减少乘客换乘次数可有效提升城市轨道交通运输服务水平,本文结合线网客流特征提出跨线方案备选方法,构建跨线列车开行方案优化模型。首先,根据研制的出行路径推算系统分析线网全日各类客流量及比例,确定线网跨线列车的跨线次数,获得备选跨线长交路集;其次,以乘客换乘次数最小化为目标构建满足运营基本条件和跨线能力约束的跨线列车开行方案模型,并嵌套基于发车频率客流分配的改进遗传算法求解,获得本线交路和跨线交路的开行频率;最后,以某城市轨道交通线网为例,研究跨线列车开行方案的优化效果。结果表明:在换乘站开行跨线长交路,全网换乘乘客总换乘次数减少了2.02%~5.97%,降低乘客换乘时间和线网换乘系数,同时直达客流增加了1.58%~4.58%;跨线长交路在衔接线路上承担小交路以补充断面运力,减少本线交路列车的开行频率,可减少线网的上线列车总数和列车正线走行公里数;此外,当换乘站的换乘客流量较高时,跨线长交路因衔接线路上本线交路列车开行频率较高而无法实施,且随着跨线交路开行频率接近跨线能力上限,运营服务水平的提升效果呈逐渐降低趋势。

收费及奖励策略对通勤出行方式选择的影响分析

为研究收费及奖励两种交通需求管理策略对居民通勤出行方式选择的调控作用,以杭州市主城区为例,通过SP(Stated Preference)和RP(Revealed Preference)调查私家车通勤出行者在停车收费及非小汽车出行奖励情景下的方式选择意向,利用非集计理论分别建立停车收费和出行奖励单独实施和共同实施下通勤出行方式选择的巢式Logit(NL)模型,研究停车收费及出行奖励对私家车通勤出行者出行方式选择的单独影响及共同影响效果,并分析两类策略的弹性。研究结果表明:停车收费和非小汽车出行奖励均能减少私家车出行比例,增加公交和地铁等公共交通出行比例;当停车收费达到一定费率时,才能有效减少私家车出行,适当的奖励能有效激励出行者转向其他出行方式,若同时实施收费和奖励策略,将表现为收费为主和奖励为辅的联合作用;在3类场景中,收入均是影响出行方式选择的重要因素,收入越高,继续选择私家车出行的可能性越大;单独收费场景下,停车收费弹性随费率增加而增大,且费率较低时缺乏弹性,出行奖励弹性随奖励金额增加而先增大后减小,小额奖励即有弹性。

考虑空间需求不均的模块化公交线路运行方案优化研究

传统固定容量公交车辆难以应对公交线路上空间分布不均的需求。为解决这一问题,引入模块化公交车辆,通过编组与解编的方式实现车队在运行中的容量动态变化,以更贴合需求在空间上的变化。基于运行时空图重构的方法,建立描述模块化公交线路运行方案的优化模型,所构建模型为混合整数非线性规划(Mixed Integer-Nonlinear Program,MINLP)模型,模型决策变量包括车队的运行方案以及模块化公交单元的运行方案。为方便求解,采用时间离散化手段,通过一系列方法将所构建的MINLP模型转化为易于求解的混合整数线性规划(Mixed Integer-Linear Program,MILP)模型,并基于成都市真实公交线路与乘客需求数据进行案例分析。实验结果表明,与传统容量固定式公交相比,模块化公交的应用能够使乘客成本降低11.44%,运营成本降低31.35%,综合降低20.32%的系统总成本。敏感性分析实验结果分析了系统供给与需求变化对系统成本产生的影响。

考虑订单优先级带时间窗的多车型开放式车辆路径问题研究

针对物流运输中具有优先级的配送订单情况以及新能源车辆、燃油车辆混合配送的车辆路径优化问题,考虑车辆可行驶区域限制、车辆载重量、客户送货时间需求、新能源车辆充电约束,构建以碳排放成本、货运成本和时间窗惩罚成本总和最小的目标函数。根据订单优先配送特征设计带有优先策略的粒子群算法求解问题,并基于条件采用适应性算例进行实验,验证算法对考虑优先级订单序列带有时间窗的多车型开放式车辆路径问题的有效性。

基于多智能体元强化学习的危险品运输路径优化

针对危险品运输车辆路径优化问题,本研究基于运输公司多车辆服务全部客户的现实需求,通过多智能体系统提高车辆之间的协同效率,以不同权重的旅行时间和安全风险最小为运输路径优化目标,同时兼顾时间窗、载货量等约束,构建多智能体强化学习模型,并采用元强化学习方法,建立更具泛化能力的元模型。将不同权重下的危险品运输问题抽象为带时间窗的多车辆多行程运输路径优化子任务,利用深度网络模型的不同嵌入层刻画子任务的高维特征。通过有效结合元学习Reptile算法思想与滚动基线方法训练元模型,前者增强了优化方法对不同子任务的适应性,后者则通过贪婪地选择具有最大概率的动作提高了优化方法在各子任务求解计算中的灵活性。实验结果表明:本文采用的多智能体元强化学习方法相对于迁移强化学习方法,在非支配点数量和超体积两个指标上,分别提升了12%和22%,说明其更接近帕累托最优解;而在不同解码方法中,集束采样方法更具优势。