EFRI ARES-CI Embedded Distributed Simulation for Transportation System Management

项目编号：0735991EFRI ARES-CI交通系统管理嵌入式分布式仿真从高速公路上的交通事故到飓风或人为灾害等重大灾难，都需要复杂的系统重构过程。这些系统的复杂性要求利用复杂的在线决策支持工具来有效地管理重大危机。该项目解决了有关有效实现由网络基础设施实现的自主可重构工程系统的研究挑战。该研究结果已应用于地面交通系统的管理。该研究探索了一种新的、具有潜在变革性的实时系统仿真和系统管理方法，即结合实时信息的ad-hoc分布式仿真，可快速生成用于系统管理和优化的未来系统状态的可靠预测。Ad-hoc分布式仿真的特点是在连续运行的实时分布式仿真环境中，自动仿真器相互交互和实时数据的动态集合。它们结合了两个传统上独立领域的概念：分布式仿真和复制仿真运行的输出分析。计算研究的重点是创建有效的临时分布式模拟所需的基础理论和算法，考虑有和没有路边基础设施的范式。分析研究侧重于在线统计方法来处理从环境中收集的数据，以及自适应学习技术来实时校准模拟预测。实验研究侧重于ad-hoc分布式仿真系统的工程方面，并将在仪表车辆和启用无线的路边计算机上部署原型系统。这项对地面交通领域的研究有可能为研究、分析和缓解系统弱点提供更好的工具和技术，为更好的应急响应计划提供决策支持工具，并在紧急情况下提供更好的态势感知可视化。临时分布式仿真概念也广泛适用于除运输之外的各种系统，例如制造和供应链管理。教育外展活动将通过每年夏天让一到两名教师参与项目，并邀请少数族裔本科生作为实习生，增加工程和计算机领域代表性不足群体的参与。研究人员还将把研究活动和结果纳入研究生水平的课程，统计方面的进展将纳入博士水平的模拟课程。

PI: Hunter, Michael P.Institution: Georgia Institute of TechnologyProposal Number: 0735991EFRI ARES-CI Embedded Distributed Simulation for Transportation System ManagementUnexpected events ranging from traffic accidents on freeways to major catastrophes such as hurricanes or man-made disasters require sophisticated system reconfiguration processes. The complexity of such systems mandates the exploitation of sophisticated on-line decision support tools to effectively manage major crises. This project addresses research challenges concerning the effective realization of autonomously reconfigurable engineering systems enabled by cyberinfrasturcture. The research is applied to the management of surface transportation systems. The research explores a new, potentially transformative approach to real-time system simulation and system management, namely, ad-hoc distributed simulations incorporating real-time information rapidly producing reliable forecasts of future system states for use in system management and optimization. Ad-hoc distributed simulations feature dynamic collections of autonomous simulators interacting with each other and real-time data in a continuously running real-time distributed simulation environment. They combine concepts from two traditionally separate domains: distributed simulation and output analysis of replicate simulation runs. Computational research focuses on the underlying theory and algorithms necessary to create effective ad hoc distributed simulations, considering paradigms with and without roadside infrastructure. Analytical research focuses on on-line statistical methods to process data collected from the environment as well as adaptive learning techniques to calibrate simulation predictions in real-time. Experimental research focuses on aspects concerning the engineering of ad-hoc distributed simulation systems and will deploy a prototype system on instrumented vehicles and wireless enabled roadside computers.This research into the surface transportation arena has the potential to lead to better tools and techniques for the study, analysis, and mitigation of system weaknesses, decision support tools for better emergency response planning, and visualizations offering better situational awareness during emergencies. The ad-hoc distributed simulation concept also has broad applicability to a wide variety of systems in addition to transportation, such as manufacturing and supply chain management. Educational outreach activities will increase participation of underrepresented groups in engineering and computing by involving one or two teachers in the project each summer and hosting minority undergraduate students as interns. The investigators will also incorporate research activities and results in graduate-level curriculum and the statistical advancements will be incorporated into a Ph.D.-level simulation course.