报告题目：Learning to Schedule Multi-NUMA Virtual Machines via Reinforcement Learning
报 告 人：王祥丰 教授（华东师范大学）
报告时间：2021年11月8日 下午2:00开始
报告地点：腾讯会议 366 842 021 会议密码：2021
报告摘要：With the rapid development of cloud computing, the importance of dynamic virtual machine scheduling is increasing. Existing works formulate the VM scheduling as a bin-packing problem and design greedy methods to solve it. However, cloud service providers widely adopt multi-NUMA architecture servers in recent years, and existing methods do not consider the architecture. This paper formulates the multi-NUMA VM scheduling into a novel structured combinatorial optimization and transforms it into a reinforcement learning problem. We propose a reinforcement learning algorithm called SchedRL with a delta reward scheme and an episodic guided sampling strategy to solve the problem efficiently. Evaluating on a public dataset of Azure under two different scenarios, our SchedRL outperforms FirstFit and BestFit on the fulfill number and allocation rate.
报告人简介：王祥丰，华东师范大学计算机科学与技术学院副教授，2009年和2014年分别获得南京大学学士和博士学位；攻读博士学位期间，获得国家留学基金委资助赴美国明尼苏达大学联合培养。毕业后，加入华东师范大学计算机科学与技术学院，主要研究方向是多智能体强化学习、分布式优化、可信机器学习等。已在IEEE Transactions on Pattern Analysis and Machine Intelligence、IEEE Transactions on Cybernetics、IEEE Transactions on Signal Processing、IEEE Transactions on Medical Imaging、《软件学报》等人工智能国际权威期刊以及CVPR、IJCAI、AAAI、AAMAS、UAI、ICMR、ICASSP等人工智能国际权威会议发表论文30余篇；已在Mathematical Programming、Mathematics of Operations Research、SIAM Journal on Scientific Computing等运筹学国际权威期刊发表论文。目前担任上海市运筹学会青年委员、上海市计算机学会青工委委员。
