港口吞吐量概率预测模型研究

为了模拟港口重叠腹地货运流向情况,并对新建港口吞吐量进行合理预测,引入了概率交通方式预测模型并对其进行改进,且在此基础上建立了负指数货运量概率预测模型。将该模型应用于徐州地区新建港区的货运量分析,预测结果表明新模型具有一定的实用性,可以为港口规划和建设规模的确定提供重要参考。     

Identification and interpretation of spatial–temporal mismatch between taxi demand and supply using global positioning system data

Taxis provide a type of on-demand mobility service in urban transportation. There are growing concerns regarding the level of service of taxis in satisfying residents’ mobility demand. However, the mismatch between demand and supply of taxi services has not been resolved even with the introduction of app-based vehicle services. In this study, three indicators are constructed using the Shanghai taxi global positioning system (GPS) dataset to reveal the spatial–temporal mismatch pattern. The issue is further analyzed via a case study, in which seven regions with frequent mismatch patterns are identified in the time period from 21:00 to 22:00. Then, a multinomial logistic regression model is employed to identify the demographic and built-environment factors contributing to the taxi mismatch problem. The study results show that population density, residential areas, the number of points of interest, and road density have significant relationships with the taxi undersupply. In contrast, the areas of commercial lands, as well as the other two transportation-related factors, that is, the number of bus stops and the distance to nearest subway station, are observed to be statistically significant for both “oversupply” and “undersupply” conditions. This study provides valuable insights for identifying mismatch patterns and interpreting the mismatch problem as a function of spatially explicit factors that are of great use for urban governance, especially in the improvement of taxi efficiency, taxi management, and urban planning.     

铁路客运服务需求的模糊聚类分析

为了掌握客运服务各要素在铁路客运服务需求中的重要程度,提出按其重要程度归类的ISODATA模糊聚类方法。根据近期客运市场调查数据,确定客运服务要素的4个特征指标。运用ISODATA模糊聚类方法对21个客运服务要素进行聚类,并对聚类效果进行评价。把21个服务要素按其重要程度分成5类,由21个客运服务要素归类结果,分析确定要素在客运服务市场中的地位。     

自动全站仪隧道围岩变形非接触监测及分析预报系统研究

论述用自动全站仪结合计算机组成隧道围岩变形量测及分析系统的原理、功能及其开发应用,提出全站仪双站自由设站三维坐标非接触量测、单站独立坐标测线、双站独立坐标测线法等围岩变形量测的理论和方法,并建立相应的数学平差模型,使围岩变形非接触监测具有更好的可靠性和精度。简述全站仪机载软件、数据后处理软件的设计和开发,在机载软件控制下,无需进行对中、量仪高,全站仪可自动完成对目标点的监测,由计算机进行所有的数据处理、回归分析和预报,为隧道施工提供及时的信息反馈。     

大型造船企业成本核算系统的设计与实现

本文根据中国船舶工业集团公司所属四大造船企业的特点，对造船成本核算系统具有共性的需求进行了研究分析，设计和实现了对工资、辅助生产、制造费用、制造管理费等费用的分配，对系统软件结构、数据间的关联和业务处理时序进行了论述。     

跨国公司—当今国际海运市场需求的主要来源

当今以国际贸易来调剂或弥补国与国间各自资源稀缺状况的手段已不能完全适应世界经济发展的需要，对外直接投资及其组织形态——跨国公司能使资源在全世界范围内得到有效配置。跨国公司生产和经营特性决定了其对国际运输，特别对国际海运的作用，也决定了国际贸易的构成。本文在此认识上，分析了国际海运与国际贸易、跨国公司之间的关系；论述了跨国公司对国际海运需求的主要来源，归纳了由此需求而产生的对国际海运发展模式和趋势的影响。     

基于OD反推技术的城市新建高等级道路交通量预测方法--以镇江市长江隧道为例

随着新一轮路网调整工作在各个城市开展,隧道、高架类型的城市高等级道路建设数量不断增多,由于这些高等级道路的建设造价较大,因此在初期决策阶段需要对这些道路的交通量进行快速、准确的预测以便为决策者及规划者提供直观的评价依据。然而传统的“四阶段”方法耗时耗力,无法满足该项工作需要。鉴于此,结合镇江市长江隧道交通量预测,对基于OD反推技术的城市新建高等级道路交通量预测方法进行了介绍,并对其预测结果与传统“四阶段”法进行了对比分析。     

基于GM-BP神经网络的交通量预测

针对城市道路交通系统的复杂性和随机性,应用灰色理论和神经网络知识,建立了基于灰色理论和BP神经网络的城市道路交通量GM-BP神经网络预测模型.随后运用该预测模型对城市道路的交通量进行预测,预测结果表明：GM-BP神经网络预测模型所得预测结果平均相对误差为1.17%,与单一的灰色新陈代谢预测模型相比具有预测精度高的优点.     

The Added Value of Accounting For Users’ Flexibility and Information on the Potential of a Station-Based One-Way Car-Sharing System: An Application in Lisbon,Portugal

Car-sharing systems are an alternative to private transportation whereby a person may use an automobile without having to own the vehicle. The classical systems in Europe are organized in stations scattered around the city where a person may pick up a vehicle and afterward return it to the same station (round trip). Allowing a person to drop off the vehicle at any station, called one-way system, poses a significant logistics problem because it creates a significant stock imbalance at the stations, which means that there will be times when users will not have a vehicle available for their trip. Previous mathematical programming formulations have tried to overcome this limitation by optimizing trip selection and station location in a city in order to capture the best trips for balancing the system. But there was one main limitation: The users were assumed to be inflexible with respect to their choice of a station, and held to use only the one closest to their origin and destination. If the user is willing to use the second or even the third closest station the user could benefit from using real-time information on vehicle stocks at each station and be able to select the one with available capacity. In this article we extend a previous model for trip selection and station location that takes that aspect into account by considering more vehicle pick-up and drop-off station options and then apply it to a trip origin–destination matrix from the Lisbon region in Portugal. Through the extended formulation we were able to conclude that user flexibility allied with having information on vehicle stocks increases the profit of the company, as people will go directly to a station with a vehicle available, thus making the use of the fleet more efficient. Observing the size of the stations resulting from the model, we also concluded that the effect of information is enhanced by large car-sharing systems consisting of many small stations.     

基于站点实时关联度的短时公交客流预测方法

为探究公交站点之间的关联度并对公交客流进行更精准的实时预测，本文提出基于 Attention的交通预测核心算法(Traffic Forecast Model Based Attention，TFMA)，结合数据预处理和 站点信息编码完成基于站点实时关联度的短时公交客流预测方法。该方法首先创新性地提出了 站点实时关联度，可实现对目标站点客流量更精准的预测；其次，在公交站点的编码信息中融入 线路站点信息、客流变化率、天气、日期等关联因素；接着，该方法依靠Attention机制计算站点实 时关联度；核心算法中使用multi-headed机制、增加通道和残差连接进一步提升预测能力；最后， 以苏州市公交数据进行验证。结果显示：在准确率上，对比多元线性回归的53.8%、GRU(Gated Recurrent Unit)的66.9%和LightGBM(Light Gradient Boosting Machine)的81.2%，本文提出的基于 站点实时关联度的短时公交客流预测方法的准确率在90%以上，表明该方法具备优秀的短时公 交客流预测能力。