基于遗传算法的BP神经网络短时交通流预测

对BP神经网络模型、方法及遗传算法的基本原理进行了分析,将遗传算法与标准BP神经网络算法相结合构建了基于遗传算法的神经网络算法;建立了基于BP神经网络的短时交通流预测模型,对BP神经网络算法及基于遗传算法的BP神经网络算法进行设计,并将其应用于短时交通流预测模型的仿真计算,仿真结果表明基于遗传算法的BP神经网络算法具有更快的计算速度及更好的适应能力,该方法可以较好地应用于短时交通流预测。     

混凝土结构钢筋锈蚀率BP神经网络预测模型

文章基于对钢筋锈蚀率与锈胀裂缝关系的分析,建立了根据锈胀裂缝宽度预测钢筋锈蚀率的BP神经网络模型,并将该模型的预测结果与采用线性拟合法预测的结果进行了比较。结果表明：用BP神经网络模型可以较准确、简便、快速地预测钢筋混凝土结构的钢筋锈蚀率。     

基于PCA-BP神经网络的铁路客运量预测模型研究

本文分析铁路客运量影响因素,利用主成分分析(PCA)消除原始铁路客运量影响因素之间的相关性,将主成分分析结果作为BP神经网络的输入,并通过增加动量项、输入数据处理、调整学习速率优化BP神经网络,提出基于PCA-BP神经网络的铁路客运量预测模型。实例研究表明,与BP神经网络相比,PCABP神经网络能有效提高铁路客运量预测精度。     

灰色BP网络串联式组合模型在铁路客货运输量预测中的应用

针对铁路客货运输量发展趋势的研究,建立一种基于灰色理论和BP神经网络的串联式组合预测模型。该模型首先用同一组数据序列建立不同参数的灰色方程,然后用各灰色方程分别预测,最后将各灰色方程预测的结果进行BP神经网络非线性组合,形成串联式组合预测模型。对湖南省铁路客货运量进行分析预测,结果表明:该组合模型预测的准确性高于单独使用灰色模型的准确性,是一种可靠有效的预测方法。     

隧道洞室地基稳定性双阶段多策略粒子群BP网络模型研究

影响隧道洞室地基稳定性的因素众多,这些因素与隧道洞室地基稳定性之间存在着复杂的非线性关系,并且常规的方法很难描述这种复杂的关系。文章提出了一种双阶段多策略粒子群算法(DMPSO)优化的BP神经网络隧道洞室地基稳定性评价模型。粒子群算法具有全局优化能力强、搜索效率高等特点,算法改进后使这些特点更加突出。BP算法有很强的非线性映射能力、泛化能力,但也有收敛速度慢,容易陷入局部最优等缺陷。采用双阶段多策略粒子群算法(DMPSO)搜索BP模型的权值和阈值,弥补了BP模型的缺陷,提高了其预测的准确度。文章以重庆小什字车站洞室地基为例,证明了双阶段多策略粒子群算法优化的BP神经网络模型(DMPSO-BP)的可行性,并且该模型比模糊神经网络和粒子群优化的BP神经网络(PSO-BP)模型有更好的预测精度。     

BP神经网络算法下青海公路货运周转量预测

随着中西部社会化建设的迅猛发展,中西部许多地区的运输发展直接影响着丝绸之路经济带的建设,我们通过研究预测未来的需求制定相关措施来助力丝绸之路的经济发展,论文将青海省公路货运量与BP神经网络模型相结合。利用MATLAB对BP神经网络模型预测未来公路货物运输相关数据。将预测结果作为依据来指导下一步的相关工作,对研究青海公路运输有一定参考意义。     

基于人工蜂群优化小波神经网络的隧道沉降预测

小波神经网络存在收敛速度较慢、易陷入局部最优的缺陷,而人工蜂群算法收敛速度快且同时具有局部和全局搜索的能力。文章利用人工蜂群算法对小波神经网络进行优化,形成人工蜂群小波神经网络,并将其应用于地铁隧道沉降预测;并以深圳地铁10号线为例,将该模型的预测结果与BP神经网络、小波神经网络进行对比分析。结果表明,人工蜂群小波神经网络较其它两种模型的预测精度更高,预测结果更稳定。     

基于BP神经网络的川气东送管道土壤腐蚀预测

通过现场埋片和土壤理化性质试验确定了影响川气东送管道腐蚀速率的主要土壤因素,对39个试验点的X70钢质试片的腐蚀数据进行分析,构建了X70钢材土壤腐蚀预测模型,利用该模型在BP神经网络中训练、模拟,并运用MATLAB软件对神经网络进行编程,将预测结果与现场X70埋片腐蚀实验结果对比。结果表明:运用BP神经网络可以建立稳定性好的土壤腐蚀预测模型,预测川气东送管道X70钢材在土壤中的腐蚀速率的准确度达到90%以上。     

基于SVM-BP神经网络组合模型的高速公路出口流量预测

为了给高速公路交通流精准预测提供更准确的方法,本文利用济南西高速公路出口早晚高峰流量数据,采用SVM-BP神经网络组合模型进行短时交通流预测,并对单一的SVM（支持向量机）模型、BP神经网络模型和组合模型的预测精度进行比较和实证分析。当样本数量小于或等于120时,结果表明：(1)误差对比：当样本数量大于22时,由于预测集与训练集数据分布本身存在差异且SVM模型训练完成后过于复杂导致三种模型的误差逐渐变大。(2)预测精度：组合模型>BP神经网络>SVM,组合模型的平均绝对误差（MAE）提高了6.85%,远高于其他单一模型,验证了组合模型的有效性。     

基于数据分析的铁路列车到达晚点预测模型

准确估计列车的实际晚点时间对于列车运行图制定、列车调度员决策具有重要意义。为此,以荷兰铁路列车运行实际数据为研究基础,根据列车实际运行数据选择自变量,以列车的实际晚点时间为因变量,建立列车晚点的随机森林预测模型,并以BP神经网络模型作为对比,结果表明随机森林模型预测误差稳定,预测精度较高,比传统的神经网络方法取得了更好的预测效果。     

共享单车停放需求影响因素分析和预测研究

为了解决城市共享单车的乱停乱放问题,本文基于北京市的共享单车出行大数据,提出了共享单车停放需求预测的多项Logit模型。首先分析了单车停放需求的影响因素,然后选取了时间、空间及天气方面的12个因素为自变量,通过Wald检验分析了这些因素与停放需求的相关性和显著性,基于多项Logit模型建立了共享单车的停放需求预测模型。结果表明:工作日、时段、商业区、所临道路类型、临近轨交站、高温、下雨、以及风力等级与共享单车停放需求显著相关;构建的预测模型总体预测准确率为77.5%,其中对出现频率最高的低停放需求预测准确率高达86.49%。     

A study on smart parking guidance algorithm

Parking problem becomes one of major issues in the city transportation management since the spatial resource of a city is limited and the parking cost is expensive. Lots of cars on the road should spend unnecessary time and consume energy during searching for parking due to limited parking space. To cope with these limitations and give more intelligent solutions to drivers in the selection of parking facility, this study proposes a smart parking guidance algorithm. The proposed algorithm supports drivers to find the most appropriate parking facility considering real-time status of parking facilities in a city. To suggest the most suitable parking facility, several factors such as driving distance to the guided parking facility, walking distance from the guided parking facility to destination, expected parking cost, and traffic congestion due to parking guidance, are considered in the proposed algorithm. To evaluate the effectiveness of the proposed algorithm, simulation tests have been carried out. The proposed algorithm helps to maximize the utilization of space resources of a city, and reduce unnecessary energy consumption and CO₂ emission of wandering cars since it is designed to control the utilization of parking facility efficiently and reduce traffic congestion due to parking space search.     

Effect of parking information on travelers' knowledge and behavior

In many urban centers the demand for parking increases sharply before Christmas mainly due to shopping activity — causing parking congestion. One way to ameliorate such congestion is by disseminating parking information. Informed drivers may divert to relatively under-utilized parking facilities relieving the pressure on congested facilities. The City of Nottingham in England tested a real-time parking information system designed to alleviate congestion in the City Center parking facilities. Real-time information was disseminated through the radio, while historical information regarding parking locations was disseminated through newspaper advertisements and leaflets. The objective of this study is to assess impacts of the parking information system on travelers' knowledge and decisions.Survey research was used to understand traveler response. Respondents' levels of knowledge regarding car parks were analyzed by developing Poisson regression models. Drivers were more likely to have greater knowledge of city center car parks if they used several information sources (radio broadcasts, newspaper advertisements or leaflets and word-of-mouth), were active seekers of parking information, and searched for parking rather than going directly to a parking facility. In addition, the study of behavior showed that drivers were more inclined to use the relatively under-utilized Park-and-Ride facilities instead of the city center car parks if they received parking information from Newspaper advertisements/leaflets. Overall, the parking information service in Nottingham was effective; it seems reasonable to establish such information dissemination and monitoring systems at parking facilities in other urban areas. Furthermore, to support informed travel and activity participation decisions, parking information should be integrated with traffic and transit information.     

A review of the impact of parking policy measures on travel demand

This paper reviews the empirical evidence relating to the impact of parking policy measures on the demand for parking and for travel. Disaggregate modal choice models, disaggregate parking location models and site‐specific studies of parking behaviour are examined. With regard to modal choice models, it is concluded that few studies deal adequately with parking factors, but that there is some support for the view that parking policy measures are a relatively important influence on modal choice. When parking location models are examined parking policy variables are shown to have a substantial impact on choice of parking location. With regard to site‐specific studies, the paper concludes that there is a great variation in the parking price elasticities quoted, which reflects partly the methodological problems associated with such studies. Suggestions to improve model specification are made.     

Empirical evidence on cruising for parking

The literature on car cruising is dominated by theory. We examine cruising for parking using a nation-wide random sample of car trips. We exclude employer-provided and residential parking. We focus on the Netherlands, where levels of on-street and off-street parking prices are locally the same. We demonstrate then that due to this price setting the average cruising time in the Netherlands is only 36 s per car trip. Furthermore, we show that cruising is not random. It is more common in (large) cities that receive more car trips, particularly for shopping and leisure activities. Cruising time increases with travel duration as well as with parking duration. Cruising has a distinctive pattern over the day with a peak in the morning, so the order of arrival is essential to parking. Because cruising has a spatial and time component, policies may be considered that reduce cruising time through flexible pricing of parking or improved information about vacant parking spaces.     

System dynamics of urban traffic based on its parking-related-states

The urban parking and the urban traffic systems are essential components of the overall urban transportation structure. The short-term interactions between these two systems can be highly significant and influential to their individual performance. The urban parking system, for example, can affect the searching-for-parking traffic, influencing not only overall travel speeds in the network (traffic performance), but also total driven distance (environmental conditions). In turn, the traffic performance can also affect the time drivers spend searching for parking, and ultimately, parking usage. In this study, we propose a methodology to model macroscopically such interactions and evaluate their effects on urban congestion.The model is built on a matrix describing how, over time, vehicles in an urban area transition from one parking-related state to another. With this model it is possible to estimate, based on the traffic and parking demand as well as the parking supply, the amount of vehicles searching for parking, the amount of vehicles driving on the network but not searching for parking, and the amount of vehicles parked at any given time. More importantly, it is also possible to estimate the total (or average) time spent and distance driven within each of these states. Based on that, the model can be used to design and evaluate different parking policies, to improve (or optimize) the performance of both systems.A simple numerical example is provided to show possible applications of this type. Parking policies such as increasing parking supply or shortening the maximum parking duration allowed (i.e., time controls) are tested, and their effects on traffic are estimated. The preliminary results show that time control policies can alleviate the parking-caused traffic issues without the need for providing additional parking facilities. Results also show that parking policies that intend to reduce traffic delay may, at the same time, increase the driven distance and cause negative externalities. Hence, caution must be exercised and multiple traffic metrics should be evaluated before selecting these policies.Overall, this paper shows how the system dynamics of urban traffic, based on its parking-related-states, can be used to efficiently evaluate the urban traffic and parking systems macroscopically. The proposed model can be used to estimate both, how parking availability can affect traffic performance (e.g., average time searching for parking, number of cars searching for parking); and how different traffic conditions (e.g., travel speed, density in the system) can affect drivers ability to find parking. Moreover, the proposed model can be used to study multiple strategies or scenarios for traffic operations and control, transportation planning, land use planning, or parking management and operations.     

Modeling heterogeneous parking choice behavior on university campuses

Studies on campus parking indicate more severe problems and a wider range of characteristics than commercial parking because of limited parking places, special conditions, specific policies and enclosed space on university campuses. Heterogeneous characteristics are usually ignored in analyses of campus parking behavior. In this paper, a mixed logit model is applied to analyze parking choice behavior on a campus using data collected from a stated-preference survey of Tongji University, Shanghai, China. The heterogeneity of individuals with various sociodemographic characteristics is evaluated by interaction terms and random parameters. Comparison between the proposed approach and the conditional logit model shows that the results of the mixed logit model are more interpretable because they are not limited by the independence from irrelevant alternatives assumption. Key factors that have considerable effects on campus parking choices are identified and analyzed. Important regularities are also concluded from elasticity analyses. Finally, the campus is divided into two areas according to the walking distance to a new parking lot, and the modeling results show that area-specific policies should be established because the two areas have quite distinct parking choice features.     

Bicycle parking: a systematic review of scientific literature on parking behaviour,parking preferences,and their influence on cycling and travel behaviour

ABSTRACT

Cycling is experiencing a revival in many cities. Research has focused on the determinants of cycling – in particular the role of the built environment and road infrastructure. Bicycle parking has received little attention – even though bicycles are parked most of the time. This article reviews the scientific literature on bicycle parking and identifies existing gaps in research and knowledge. The review analyses 94 peer-reviewed papers identified through a search in Scopus and Web of Science, in December 2017. The annual number of papers increased 15-fold between 1995 and 2017. Overall, the level of evidence on the importance of bicycle parking is limited. The majority of studies are based on cross-sectional data with the presence of parking as a binary independent variable. Most studies focus on bicycle parking at public transport stops and at work places. Few studies report on bicycle parking throughout cities, and hardly any on parking at residential locations. Bicycle parking supply and quality appears to be a determinant of cycling for current and potential cyclists. Our findings can serve as input for an evidence-based debate on the role of bicycle parking. For practitioners, our research supports investment in bicycle parking, but acknowledges that a proper evaluation of such initiatives needs to be conducted to increase the level of evidence.     

高速公路服务区停车位计算参数探析

文章分析了高速公路服务区规模与停车位的关系,介绍了停车位计算模型,并结合江西省高速公路服务区的实地调查情况,对交通量、驶入率、周转率及高峰率等服务区停车位计算参数进行了计算分析,得出了高速公路服务区停车位计算参数值。     

Modeling the impact of parking price policy on free-floating carsharing: Case study for Zurich,Switzerland

The research on carsharing has already shown that a non-negligible part of carsharing members give up a vehicle after joining a carsharing program, or avoid a vehicle purchase. This arguably reduces overall parking space needed. This might well be one of the most important impacts of a carsharing program on the transportation system, but also one of the least researched. The rapid diffusion of free-floating carsharing, which for its very nature might have a stronger impact on parking, makes the relationship between carsharing and parking an appealing topic for new research. This work presents a method for the investigation of this relationship using an agent-based simulation and explores the impacts of different parking prices on the demand for free-floating carsharing in the city of Zurich, Switzerland. Three levels of free-floating fleet-size in the city of Zurich coupled with three levels of parking prices were simulated. The obtained results show that free-floating vehicles are able to use parking spaces more efficiently than private vehicles. Moreover, the average parking occupancy tends to be more homogeneous with higher fleet-size of free-floating carsharing and with the increase of parking prices, thus avoiding spatial parking pressure peaks.