个性化动态路径诱导建模理论与方法研究

A novel personalized dynamic route guidance modeling theory and methods are presented in this project. The innovations are as the followed. Firstly, to solve the difficult acquisition problem of the driver's travel route choice preferences, an adaptive and self-learning model and algorithm are presented, in which the driver's travel route choice standards and preferences in different travel situations are learned and excavated by a differential perception, and the outcomes are stored and updated by a memory mechanism. Furthermore, all of the works are automatically finished in the route guidance terminal devices. Thus, it's more automatic and more advanced than the existing group questionnaire method. Secondly, in order to improve the driver's response and balance the network traffic demand, a differentiation route guidance strategy is presented, and the situational leadership theory is introduced to establish the strategy model. The core idea of the strategy and the model is to implement a personalized route guidance of "a driver, a situation, and a strategy". Compared with the traditional strategies of the system optimum or the user optimum, it's more targeted and more practical. Thirdly, with the consideration of the driver's deviation behavior in the implement process of the strategy, the traditional basic assumption of the driver strictly following the guidance route is removed, and a dynamic route guidance method based on behavior oriented and capacity constraints is presented, based on which a real-time information prompt and a route optimized dynamically based on the driver's maximum response and a congestion prevention mechanism are provided to guide the driver to the best route. Based on the above researches, the personalized route guidance aims to improve the route guidance following rate and easing the traffic congestion are achieved. And the research achievements will not only have important theoretical significance and application value to develop the urban dynamic traffic guidance system and terminal devices of route guidance, but also have a positive effect to prompt the crossing and merging of different disciplines.

本项目将突破传统基于群体问卷调查的驾驶员出行路径选择偏好提取方法，在交通诱导终端设备中通过差分感知和记忆机制获取驾驶员在不同情境状态下的出行路径选择标准和偏好，建立驾驶员个性化出行路径选择特性自学习模型和算法，解决了驾驶员出行路径选择偏好难以获取的难题；突破传统基于系统最优和用户最优的路径诱导策略，以驾驶员最大响应度和均衡路网交通负荷为目标，构建基于情境驱动的差异化动态路径诱导策略模型，实现"一人一情境一策略"的个性化诱导，更具针对性和实用性；突破已有对驾驶员严格遵循诱导路径的假设，从满足个性需求和预防拥挤转移的角度，提出基于容量约束和行为导向的动态路径优化与诱导方法，将驾驶员的出行路径导向到诱导路径上来，实现提高诱导服从率、防止交通拥挤的个性化诱导。研究成果不仅对开发城市动态交通诱导系统和交通诱导终端设备具有重要的理论意义和应用价值，而且对促进多学科交叉与融合有积极作用。

项目组主要从以下三个方面展开理论研究和技术开发，具体如下：. 1）以交通诱导终端设备为平台，研究驾驶员个性化出行路径选择特性的自学习模型和算法，以描述不同情境状态下驾驶员“最优”出行路径的选择标准和偏好，从而提出驾驶员个性化出行路径选择特性的感知和学习方法，为个性化动态路径诱导提供依据。. 2）以交通信息中心为平台，研究基于情境领导理论的差异化动态路径诱导策略研究，分析交通情境状态、驾驶员的出行路径选择行为、诱导策略三种之间博弈演化机理和关联关系，从而建立基于情境驱动的差异化动态路径诱导策略，从而为个性化动态路径诱导提供指导。. 3）以集群计算中心为平台，研究基于容量约束和行为导向的动态路径优化与诱导方法，从提高诱导服从率入手，提出满足驾驶员个性需求和防止交通拥挤转移的最优路径搜索机制和优化方法，将驾驶员的行驶路径导向到交通诱导系统下发的诱导路径上，引导驾驶员避开拥挤走最佳行驶路线。. 通过本项目的研究，实现了对不同情境状态下驾驶员出行路径选择的影响因素及其敏感性，并利用交通诱导终端设备来感知、学习驾驶员在不同情境状态下的路径选择偏好；分析了不同情境状态下道路交通环境、驾驶员、交通诱导系统之间的相互作用关系，明确了诱导条件下驾驶员路径选择行为的演化博弈机理，建立了基于情境驱动的差异化动态路径诱导策略，为个性化交通诱导提供指导；从满足个性化出行服务和防止交通拥挤转移的角度构建基于容量约束和行为导向的动态路径优化与诱导方法，为开发动态交通诱导系统和交通诱导终端设备提供新的思路和理论依据。. 项目的研究成果对提高驾驶员的诱导服从率、缓解城市交通拥挤，对促进交通诱导终端设备、动态交通诱导系统的开发与应用提供理论基础和技术支持。项目取得的具体成果如下：. 1）撰写项目研究报告1套；. 2）在国际、国内期刊上发表学术论文11篇；. 3）申请国家发明专利13项；. 4）申请计算机软件著作权2项。