題目:A distribution uncertainty model and its robust optimization approach
報告人:王然副教授(南京航空航天大學計算機科學與技術學院)
時間:2020年9月16日(周三)10:00
地點:明故宮校區A18-209會議室
主辦單位:科協、航空學院、航空航天交叉研究院、青年教授學術交流聯誼會
報告內容摘要:
In this talk, I will present our recent progress on how to handle decision-making optimization problems with uncertain variables. Normally, solving an optimization problem with random variables in constraints or objective function is cumbersome or even computational intractable. To better capture the fluctuant nature of random variables, we propose a flexible distribution uncertainty model. Specifically, we extract reference distributions as useful references and allow the actual distributions of the random variables to vary around their references. Such a model allows convenient handling of volatile variables as long as their profiles are not too intensely different from the predictions or empirical knowledge. We further develop chance-constraint approximation and robust optimization approaches based on the proposed uncertainty model to firstly transform and then solve the problem. The proposed method is a universal method that can be conveniently extended to handle many other decision-making optimization problems with uncertainties. We summarize the merit of the proposed framework and its applications in modeling and solving optimization problems under various scenarios in different networks (smart grid, wireless sensor networks, data centers).
報告人簡介:
王然,南京航空航天大學計算機科學與技術學院副教授。2011年7月本科畢業于哈爾濱工業大學實驗學院電子信息工程專業,2016年4月博士畢業于新加坡南洋理工大學計算機科學與工程學院。2011年獲“南洋工程博士獎”,2017年入選江蘇省“雙創博士”。主持國家自然科學基金、江蘇省自然科學基金、CCF-騰訊犀牛鳥基金、中國博士后特別資助等10余項科研項目;發表科研論文49篇,其中期刊論文22篇(均為CCF推薦或中科院2區以上的SCI),國際會議論文27篇(均為CCF推薦或EI),Google Scholar他引700余次;第1作者在Springer上出版英文專著1部,出版江蘇省十二五規劃教材1部(清華大學出版社);申請發明專利19項(授權2項)。所指導的本科生獲計算機學院優秀畢業論文,赴南洋理工大學直博;所指導的碩士生獲國家獎學金,校優秀畢業生等榮譽,研究方向包括智能電網、能源互聯網、最優化理論等。