NSF/USDOT Partnership for Exploratory Research - ICSST: Cooperative Traffic Management and Route Guidance: A Multi-agent Based Approach

摘要：本研究将探索一种基于多智能体协商的交通管理和路径诱导方法。在被称为协同运输管理和路线诱导系统(CTMRGS)的多智能体系统中，将IRANS(车载路径和导航系统)和ISPs(信息服务提供商)表示为智能体。这一想法是为了让所有的司机和他们的爱尔兰人有能力选择、进行一次安全、高效、划算的旅行并支付费用。每个IRANS代理商都会与互联网服务提供商代理商就行程进行协商，以便对首选路线进行最小程度的更改，从而确保司机会遵守规定。网络服务提供商将预期的道路使用情况告知网络管理人员，以换取将在服务提供商的协商模型中使用的当前交通状况。拟议的谈判基于一种新的基于原则性谈判的方法，与文献中讨论的其他多智能体谈判和讨价还价方法(例如合同网)有很大不同。因此，协商并不完全基于个体主体的效用最大化准则。在协商路径之前，代理交换它们的客观标准，以便在路径生成和改变功能中包括彼此的目标。拟议的多代理系统符合ATIS6综合运输管理/路线指南市场一揽子计划，特别侧重于克服目前ATIS6一揽子计划中存在的司机不遵守规定的障碍。

Abstract: This research effort will explore a multi-agent negotiation-based approach to traffic management and route guidance. In the proposed multi-agent system called Cooperative Transportation Management and Route Guidance System (CTMRGS), IRANS (In-vehicle Routing and Navigation Systems) and ISPs (Information Service Providers) are represented as agents. The idea is to allow all drivers and their IRANS to have the ability to choose, make, and pay for a safe, efficient, cost-effective trip. Each IRANS agent negotiates with the ISP agents about the trip for a minimal change in the preferred routes and thus ensures that the driver will comply with it. The ISP agents inform network managers with the anticipated road usages in exchange for the current traffic conditions to be used in the ISP's negotiation model. The proposed negotiation is based on a new principled negotiation based approach that is significantly different from other multi-agent negotiation and bargaining approaches (e.g., contract-net) discussed in the literature. Therefore negotiation is not entirely based on individual agent's utility maximization criteria. The agents exchange their objective criteria before they negotiate for routes so that each other's objectives are included in the route generation and alteration functions. The proposed multi-agent system is compliant with ATIS6 market package for Integrated Transportation Management/Route Guidance with specific focus on overcoming drivers' non-compliance hurdles present in the current ATIS6 package.