Research On Dynamic Route Choice Models And Methods Under Multi-sources Traffic Information

Nowadays,with fast economy development and urbanization speed,traffic congestion and safety problems have spread worldwide cities.They lead to huge economic losses as well as environmental degradation and energy consumption,and affected people's daily travel activities seriously.According to statistics,there are 15 cities with one billion losses every day caused by traffic congestion,and this may lead to 5 percent to 8 percent of national GDP one year. So how to solve this difficult livelihood issue is excepted in the whole country.However, there are two roads before us,building roads and employ Intelligent Transportation Systems(ITS).No doubt improve existing roads' efficiency is the most perfect way,when urban land resources are limited.As the core of ITS,traffic flow guidance system(TFGS) has become more significant in urban traffic managements.Based on high-tech means,TFGS picks up real-time traffic data to guidance information that include traffic incidents,travel time and optimal path.So drivers can avoid crowded sections through central or distributed guidance type,then relieve traffic congestion by made traffic flow balanced distribution in the network.As a key technology of TFGS, dynamic route choice is a general matter to every traveller. And route choice is essentially a task that choose a suitable impedance function under complex and multi-sources information,then realize the flow optimization dynamically on the road network,and calculate an optimal path depend on several conditions through right algorithm.The process optimize travellers reasonable travel at the micro and balance traffic flow's distribution at the macro.Supported by National High-Tech Research and Development Program(863 Program)and Kunming Public Security Bureau traffic police detachment program,this thesis is aimed mainly at key technologies of dynamic route choice models and algorithms under multi-sources. Specifically, it lays emphasis on the research of travellers' travel behaviour under multi-sources traffic information, dynamic route choice models and algorithms before travelling, public transportation route choice models and algorithms affected by transfer activities and adaptive route choice models and algorithm on travel. And this thesis consists of six chapters.The first chapter introduced the research background of the thesis,the research status at home and abroad,research targets and ideas. Chapter two to Chapter five are the core content of this thesis,they have studied on dynamic route choice models and algorithm against travellers' travel behaviours under multi-sources traffic information and different requirements during different travel periods and travel styles.And Chapter six has generalized summaries,key innovations of the whole thesis and made a forward-looking for further work.The main achievements of the thesis are as follows:1) Based on status and trends analysis of TFGS, dynamic Traffic Assignment (DTA) and dynamic route choice models, it put forward the shortages of common dynamic route choice models, and defined the targets and worth of dynamic route choice technology study, then made out study method and ideas of this study under multi-sources traffic information.2) Starting from the contents and extensions of multi-sources,a specific analysis was made on its concept,compositions and features,then the thesis studied on the travellers' information demand characteristics,including destinations,travel styles,route choices and starting time demands etc..And the theory, factors and assumptions for travellers' dynamic route choice behaviours under multi-sources were also proposed.Based on this study basis,this thesis analyzed travellers' route choice behaviours under different information strengths,which supported theories for route choice models.3) The key issues about route choice were analyzed that under prior information,uncertain information and multi-sources information modes.Firstly,this thesis studied on route choice models based on prospect theory when only prior information were provided,and route choice models based on cumulative prospect theory under uncertain information as well.Secondly,it put forward a multi-objectives route choice model with minimum predicting travel time as the first optimization goal and travel time reliability as constraints.Taking the abnormal accidents' influence about route choice into account,this thesis proposed a route choice model based on a real-time estimation of accidents' influence,after the analysis of the emergency route choice models based on travel time and routes complexity as well as travel time and safety.4) Firstly,it studied on the transform method of public transportation network under the influence of transfer behaviours,consisting of the method based on bus fares covertion to bus route travel time and the method based on optimal route analysis.Secondly,following the quasi-DTA principles,a bus route choice model with K optimal paths was put forward based on improved method.After that,this thesis made an extended study on bus route choice model based on Graphical Evaluation and Review Technique considering transfer time,as well as emergency bus route choice model based on time-expanded network under abnormal accidents.5) Against time-varying and timeliness characteristic of route choice on-travel,this thesis put forward a adaptive route choice model based on improved ant-method with the superiority shows in solve the optimal problems of discrete system.Firstly,it analyzed features of intermittent urban traffic flow,and then proposed a adaptive exponential smoothing method for travel time prediction under intermittent traffic flow. Secondly,with prediction travel time as route choice pheromone,this thesis established an adaptive route choice model,so it overcome the problems common route choice model has made by ant-method...