题目：大数据驱动的最短路问题研究
主讲人：李想教授
时间：2016年11月28日（周一），上午10:00
地点：A949
摘要：
In this paper, a spatial-temporal shortest path problem is proposed to search a path with lower travel time at its constituent arcs in upcoming traffic conditions. The underlying traffic big data on flow speeds is used to define the time-dependent traffic scenarios, and a forecasting-based optimization method is formulated to make the shortest path decision. First, we define a spatial-temporal road network with continuously updated speed profiles, and formulate an optimality equation for the shortest path.
In practice, since the upcoming speed profile in continuous form is unattainable, we use the past speed series collected with presetted sampling step to forecast the upcoming speed series, and reformulate a quasi-optimality equation. Then we propose a spatial-temporal Dijkstra algorithm to search the quasi-shortest path, in which a weighting process is frequently used to estimate the travel time at arcs with online calculated departure time. The relationships among the optimality on quasi-shortest path, complexity on spatial-temporal Dijkstra algorithm, forecasting deviation and sampling step are proved, which give a good picture on managing the size of traffic big data, i.e., the length of sampling step, by controlling the optimality and complexity.
Finally, case studies are conducted based on the loop detector data in Beijing road network established in January 1-31, 2014. Comparisons between forecasting-based optimization method and the traditional rolling horizon optimization method are given, which show that forecasting-based optimization method has higher optimality in non-peak hours, while rolling horizen optimization method has higher optimality in peak hours.
主讲人简介：
李想，北京化工大学教授、博士生导师。研究领域包括交通运输管理、不确定优化等，在Springer出版英文专著1部，在Transportation Research Part B、Transportation Research Part C、Transportmetrica Part B、IEEE Transactions on Fuzzy Systems、IEEE Transactions on Intelligent Transportation Systems、IEEE Transactions on Systems, Man, and Cybernetics: Systems、Information Sciences、European Journal of Operational Research等高水平期刊发表SCI收录论文60余篇，SCI引用800余次、SCI单篇最高引用90次，ESI高被引论文4篇、Elsevier Top 25热点论文4篇，荣获教育部自然科学奖一等奖、黑龙江省科技进步奖三等奖、教育部霍英东青年教师奖、钟家庆运筹学奖、运筹新人奖等，授权国家发明专利2项。研究成果得到美国工程院院士Hau L. Lee、欧洲科学院院士G.R. Chen、加拿大皇家科学院院士K.W. Hipel等同行专家的好评和实质性引用。
现任Springer丛书系列Uncertainty and Operations Research主编、国际期刊Information Sciences (SCI, IF=3.364)副主编、Transportmetrica B: Transport Dynamics (SCI/SSCI, IF=2.560)副主编、Soft Computing (SCI, IF=1.630)副主编、International Journal of General Systems (SCI, IF=1.637)编委、Sustainability (SCI/SSCI, IF=1.343)编委、《运筹与管理》(国家自然科学基金委管理学部A类期刊)编委、国际电子商务联合会（ICEB）理事、国际电子商务联合会中国分会（ICEB-CN）理事长、中国运筹学会不确定系统分会常务理事、智能计算分会常务理事、决策科学分会常务理事。曾任亚太工业工程与管理学会（APIEMS）理事、Applied Soft Computing、Journal of Ambient Intelligence and Humanized Computing、Soft Computing、International Journal of Intelligent System、Flexible Service and Manufacturing Journal等多家SCI检索期刊客座主编。
入选教育部新世纪优秀人才计划、北京市科技新星计划，主持国家自然科学基金面上项目、国家自然科学基金青年科学基金项目、教育部博士点基金项目、中国博士后科学基金一等资助项目等。
经管学院科研办
2016-11-21
