题目：随机系统的回顾性优化
报告人：王洪刚
时间：2017.7.3下午3：30-4：30
地点：bat365中文官方网站435
Talk: Retrospective Optimization for Stochastic Systems
In this talk, I will share my research work including methodology and application in simulation based optimization. I will first present R-SPLINE (Retrospective Optimization using Simplex Linear Interpolation and Neighborhood Evaluation) for integer simulation optimization problems. R-SPLINE is a stochastic optimization framework with sequential data-driven approximations and guarantees almost-sure convergence under certain conditions. I will also talk about new multi-objective optimization ideas using zigzag search for multi-criteria decision making under systems uncertainty. Numerical results based on a number of energy production and micro-grid power cases are shown to demonstrate the efficiency of the discussed algorithms.
讲座摘要：随机系统的回顾性优化
 在这次演讲中，我将分享我的研究工作，包括基于模拟优化的方法和应用。 我将首先提出R-SPLINE（使用线性插值和邻域评估的回顾性优化），用于整数模拟优化问题。 R-SPLINE是一种随机优化框架，具有依序数据驱动的近似，并保证在某些条件下几乎肯定的收敛。 我还将介绍一种新的多目标优化想法：使用之字形优化在系统不确定性下的多标准决策。我们会讨论一些能源开发，微电网管理控制的实例。通过这些例子，我们展示了所开发的随机模型和优化方法的高效性。
Bio: Honggang Wang is assistant professor in Industrial and Systems Engineering at Rutgers University. He received his Bachelor of Science degree in Power Engineering from Shanghai Jiao Tong University, Shanghai, China, in 1996, Master of Science in Manufacturing Engineering from University of Missouri-Rolla, in 2004, and Ph.D. in Operations Research from Purdue University, West Lafayette IN, in 2009. He had worked as a Postdoctoral Scholar in Energy Resources Engineering at Stanford University for two years before he joined Rutgers, NJ in 2011. Dr. Wang has won IBM faculty award 2012 and faculty award from Tracy Energy LLC for his work in petroleum and energy projects. He has won 2016 best paper award at Geothermal Research Council annual meeting, Sacramento CA, USA.
 简历：王洪刚是罗格斯大学工业与系统工程系助理教授。 他于1996年获得上海交通大学机械动力工程学士学位，2004年获密苏里大学罗拉大学制造工程科学硕士学位，2009年获普渡大学工业工程博士学位。他在斯坦福大学能源资源工程博士后工作两年，之后于2011年加入罗格斯新泽西州立大学。王博士荣获2012年IBM教授奖。 他在美国地热研究委员会2016年度会议获得最佳论文技术奖。