原文传递 Enabling Congestion Avoidance and Reduction in the Michigan-Ohio Transportation Network to Improve Supply Chain Efficiency: Freight ATIS. Progress Report, October 1, 2007.
题名: Enabling Congestion Avoidance and Reduction in the Michigan-Ohio Transportation Network to Improve Supply Chain Efficiency: Freight ATIS. Progress Report, October 1, 2007.
关键词: congestion;improve;reduct;network;transport;supply;igan;atis;chain;report
摘要: By far, the vast majority of our effort once again went toward developing static and dynamic routing algorithms that enable congestion avoidance and reduction in commercial cargo transportation networks. This goes beyond Mile-stone Number 3 in that the emphasis is not just static but both static and dynamic algorithms. The focus during the previous phase has been on developing Stochastic Dynamic Programming (SDP) based algorithms for optimal routing under Advanced Traveler Information Systems (ATIS). While SDP algorithms yield optimal policies, they are not computationally efficient. However, we need these solutions for testing and benchmarking the effectiveness of fast heuristic algorithms to be developed over the course of this multi-year project. Currently, we are adapting the more computationally efficient AO* algorithms for developing optimal policies. The process is nearly complete and preliminary results are promising. In addition, we have also extended the previous framework by incorporating more realistic nonrecurring congestion modeling and exploitation logic into the algorithms. Given that nearly 25% of all traffic congestion and about 50-60% of non-recurring delays in urban areas are attributable to incidents such as accidents (according to American Association of State Highway and Transportation Officials and National Traffic Incident Management Coalition), and that vast majority of dynamic routing algorithms reported in the literature do not exploit this information, this extension will greatly enhance the fidelity of our dynamicrouting algorithms in reducing trip completion times.
报告类型: 科技报告
检索历史
应用推荐