有序样本聚类方法在城市轨道交通运营时段划分中的应用

为合理划分轨道交通运营时段并指导其开行方案,提出一种基于有序样本聚类技术的运营时段划分方法。根据统计时段内客流数据,引入单向OD(origin-destination)概率矩阵,并给出单向OD概率矩阵的时序模型和提取方法;利用有序样本聚类方法,以最优分割法量化站间客流转移规律,求解聚类方案。最后以某一轨道交通线路为例,提取时间间隔为20 min的上行OD概率矩阵时间序列,以最优分割法进行聚类,将站间客流转移规律相近的统计时段归为一类,提出目标线路运营时段划分方案。     

基于层次分析和灰色聚类法的老龄导管架平台安全评估

为延长老龄导管架式海洋平台的使用寿命,对其进行安全评估具有重要意义.文章以渤海湾的一个老龄导管架平台为分析对象,建立了包含多重风险因素的安全评价指标体系,该体系包括三个层次:目标层,一级指标层,二级指标层.目标层包含3个一级指标:海洋环境,结构状态和人为因素.每个一级指标又分为3个二级指标.每个指标的权重通过层次分析法来计算以消除主观因素的影响.采用灰色聚类法定性和定量地评估导管架平台的失效风险.将风险评估标准分为5个等级,确定对应5个灰类的白化权函数.每个二级指标的灰色权向量由表上作业法计算得到.通过逐级计算,目标层的灰色评价值最终由灰色评价权向量单值化计算,灰色等级通过极大值原理确定.评价结果显示该老龄导管架平台失效风险等级是中等的.文中提出的安全评估方法是合理可行的.     

城市道路环境下的车载自组网分簇路由机制

车载自组网(VANET)是一种将高速移动车辆作为通信节点的Ad hoc网络,其网络拓扑结构变化迅速导致传统的Ad hoc路由技术并不适用。针对车辆节点在城市道路环境下的移动特点及分层结构给无线网络带来的易管理、易扩展等特性,提出了以公交车辆为簇头的VANET路由机制。该机制以公交车辆为簇头建立簇,其它车辆根据行驶速度和方向加入最优簇,在分簇的基础上运行路由协议。仿真实验结果表明:该机制在分组投递率和网络稳定性上均有改进。     

基于土地利用及客流特征的地铁车站分类

为研究地铁站的精细化分类问题,利用基于出行链分析的通勤出行识别方法筛选通勤客 流,结合早晚高峰的进出站客流量,识别车站的职住功能特征;基于百度地图开源平台抓取POI (Point of interest)数据,从用地功能角度进行组合归类,得到细粒度的车站周边土地利用特征。结 合以上两类特征,建立基于非监督学习K-Means++方法的地铁车站分类模型,将北京地铁307个 车站分为7类。根据其客流和周边用地特征分别识别为配套设施开发完善的典型居住型车站,具 有商业开发潜力的典型居住型车站,配置一定工作岗位的居住型车站,高度开发的典型工作型车 站,职住结合的工作型车站,旅游休闲型的车站,以及尚待开发的远郊车站。经过分析,该分类结 果与实际情况高度吻合,验证了模型的有效性,可以为城市规划及车站周边土地开发提供依据。     

基于车辆运行数据的纯电动汽车用户驾驶风格识别与驾驶能耗分析

驾驶风格是用来体现驾驶员在车辆运行状态下对车辆操作的行为特征,对用户驾驶风格进行识别与分析,有利于推进智能驾驶的发展。根据基于116 辆纯电动汽车的车辆运行数据,通过主成分分析方法与K-means 聚类算法,对用户驾驶行为进行分类分析,对驾驶风格进行了分类识别。利用XGBoost 算法构建纯电动汽车驾驶行为与能耗输入模型,利用SHAP 对模型进行解释。结果表明,将驾驶风格聚为3 类具有较好的分类效果,可分别对应冷静型、普通型与激进型;当驾驶员的驾驶风格趋向于激进型时时,车辆的驾驶能耗越高,驾驶风格激进一个层级,车辆百公里电耗增加3~4倍。当驾驶员行车时,其车速越高,油门踏板踩得越深,车辆加速度的绝对值越大,车辆的驾驶能耗越高。驾驶员的驾驶风格越激进,车辆的驾驶能耗越高。     

基于基尼系数的辽宁省现代服务业集聚与分散实证

为正确评价辽宁省现代服务业产业空间分布的均衡性,基于基尼系数分析了辽宁省现代服务业产业集聚及分散情况,研究结果表明：辽宁省新兴和现代服务业的空间集聚水平较高,传统服务业没有明显的集聚趋势;产业集中度较高的城市具有得天独厚的区位优势和根基扎实的产业基础.     

一种改进K-means模型的城市轨道交通突发事件分级方法

针对城市轨道交通事件量化分级的难题,本文提出了一种改进的 K-means聚类的突发事件分级方法.首先,从事件类型、持续时间、影响程度等方面分析各种类型事件的特征规律,提取 8个关键特征量用于聚类分析;其次,应用主成分分析法提取 4个主成分变量并提出权重系数计算方法,实现特征向量降维;提出了基于密度扫描的初始聚类中心确定方法,并将改进的 K-means聚类算法应用于地铁突发事件的分级.案例结果表明,与原始 K-means聚类方法对比,应用本文提出的改进方法聚类效果更佳.研究成果已应用于北京地铁应急指挥系统,验证了本文方法的可行性.     

基于模糊聚类的铁路智能运输系统结构分析方法研究

铁路智能运输系统(RITS)规模大、结构复杂、要素繁多,系统设计难度大.为降低系统设计复杂程度,提出了一种基于模糊聚类的RITS逻辑结构分解方法,通过建立模糊关联强度矩阵描述过程间的关联程度,在此基础上采用模糊聚类将RITS逻辑结构分解为若干相对独立的功能单元,使系统分解过程更为精细和严谨,为系统物理实现及相关研究提供了基础.     

Real time transit demand prediction capturing station interactions and impact of special events

Demand for public transportation is highly affected by passengers’ experience and the level of service provided. Thus, it is vital for transit agencies to deploy adaptive strategies to respond to changes in demand or supply in a timely manner, and prevent unwanted deterioration in service quality. In this paper, a real time prediction methodology, based on univariate and multivariate state-space models, is developed to predict the short-term passenger arrivals at transit stations. A univariate state-space model is developed at the station level. Through a hierarchical clustering algorithm with correlation distance, stations with similar demand patterns are identified. A dynamic factor model is proposed for each cluster, capturing station interdependencies through a set of common factors. Both approaches can model the effect of exogenous events (such as football games). Ensemble predictions are then obtained by combining the outputs from the two models, based on their respective accuracy. We evaluate these models using data from the 32 stations on the Central line of the London Underground (LU), operated by Transport for London (TfL). The results indicate that the proposed methodology performs well in predicting short-term station arrivals for the set of test days. For most stations, ensemble prediction has the lowest mean error, as well as the smallest range of error, and exhibits more robust performance across the test days.     

Analyzing year-to-year changes in public transport passenger behaviour using smart card data

In recent years, there has been increased interest in using completely anonymized data from smart card collection systems to better understand the behavioural habits of public transport passengers. Such an understanding can benefit urban transport planners as well as urban modelling by providing simulation models with realistic mobility patterns of transit networks. In particular, the study of temporal activities has elicited substantial interest. In this regard, a number of methods have been developed in the literature for this type of analysis, most using clustering approaches. This paper presents a two-level generative model that applies the Gaussian mixture model to regroup passengers based on their temporal habits in their public transportation usage. The strength of the proposed methodology is that it can model a continuous representation of time instead of having to employ discrete time bins. For each cluster, the approach provides typical temporal patterns that enable easy interpretation. The experiments are performed on five years of data collected by the Société de transport de l’Outaouais. The results demonstrate the efficiency of the proposed approach in identifying a reduced set of passenger clusters linked to their fare types. A five-year longitudinal analysis also shows the relative stability of public transport usage.