全校师生：

云顶集团定于2017年6月20日与6月23日举办研究生灵犀学术殿堂——李海涛报告会，现将有关事项通知如下：

1.报告会简介

报告人：李海涛

报告一：

时 间：2017年6月20日（星期二） 下午15：30

地 点：机电学院第一会议室（友谊校区航空楼B406）

主 题： Data-Driven Optimization and Prescriptive Analytics: Recent Trends, Opportunities and Applications in Supply Chains

内容简介：

With the fast growth of information technology (IT) and availability of “big-data”, data-driven decision-support is playing a more important role than ever for the success of a company or organization. In this talk, we delineate different paradigms for putting data-driven optimization in action, including rolling horizon approach, sensitivity analysis in math programming, two-stage stochastic programming, integrated simulation-optimization and stochastic dynamic programming. Focus will be put on how the Three-Pillar of Analytics: Descriptive, Predictive and Prescriptive, can be integrated to obtain efficient and high-quality data-driven solutions. We will showcase the state-of-the-art applications of data-driven optimization in supply chains, including Dell’s PC supply chain configuration for mass customization, Amazon’s supply chain network design and inventory control, precision agriculture at Monsanto, and UPS’ recent award-winning ORION project. Finally, we will foresee and discuss promising future research topics and opportunities.

报告二：

时 间：2017年6月23日（星期五） 上午9：30

地 点：机电学院第一会议室（友谊校区航空楼B406）

主 题： Approximate Dynamic Programming and Its Applications

内容简介：

Real world optimization applications in operations and supply chains are oftensequentialin nature, where decisions are madedynamicallyandadaptivelyduring the course of actions. Examples include multi-period production planning, inventory control, scheduling, routing and various resource allocation applications. While it is theoretically sound to model these problems as a dynamic programming model, or Markov decision process (MDP) for those involving uncertainty, the traditional solution approach based on Bellman’s recursion suffers the well-known curse-of-dimensionality, thus can only handle toy-size stylized problems. In this talk, we introduce the approximate dynamic programming (ADP) methodology as an effective approach to handle reasonably large MDPs. Focus will be on the three main techniques in ADP: value function approximation, forward iteration via simulation, and effective/efficient deterministic optimization methods to deal with sub-problem in each ADP iteration. We will showcase two application examples: one on large-scale stochastic resource-constrained project scheduling (SRCPSP), and the other on multi-period stochastic resource planning.

2.欢迎各学院师生前来听报告。报告会期间请关闭手机或将手机调至静音模式。

党委研究生工作部

机电学院

2017年6月19日

报告人简介

李海涛，博士，现任美国密苏里大学圣路易分校副教授、终身教授、博士生导师、商学院博士学术委员会主任。2000年毕业于北航工业外贸专业，工程学士学位，之后于2002年获得美国密西西比大学经济学硕士，200年获运营管理博士学位。李教授具有多年的优化建模和算法设计的研究经验，并长期致力于与企业，研究机构在项目调度、资源分配及供应链优化应用的合作。他曾于2004年在位于田纳西州米灵顿的美国海军人力研究科技所担任统计研究员，2005年加州帕洛阿图的惠普研究所客座研究员，2010至今任惠普研究咨询。在国际知名学术期刊发表论文20余篇，包括《European Journal of Operational Research》、《Omega》、《Journal of Scheduling》、《Annals of Operations Research》、《IEEE Transactions on Automation Science and Engineering》、《Computers and Operations Research》、《Interfaces》、《Journal of the Operational Research Society》等。李教授于2010年获得美国陆军研究所颁发的青年研究者奖、2015年密苏里大学商学院优秀研究奖。李博士拥有两项美国专利申请和多项发明公开，获得密苏里大学2015最佳发明者奖。