学术报告一
报告题目： Facility Location Games and Resource Auction Games
报 告 人： 李闽溟副教授， 香港城市大学
报告时间： 2018年5月10日 2:00-3:00PM
报告地点： 广东工业大学工学一号馆216室
简介：
李闽溟博士是香港城市大学计算机系副教授。他分别于2002及2006年在清华计算机科学与技术系取得本科和博士学位。他的主要研究兴趣是算法博弈论，组合优化，以及调度与排序问题中的算法设计与分析。
Dr. Minming Li is currently an associate professor in the Department of Computer Science, City University of Hong Kong. He received his Ph. D. and B.E. degree in the Department of Computer Science and Technology at Tsinghua University in 2006 and 2002 respectively. His research interests include algorithmic game theory, combinatorial optimization, and algorithm design and analysis for scheduling problems.
Abstract：Mechanism Design, as one of the important areas in algorithmic game theory, can be classified into two categories: with money and without money.Facility location game is one of the mostly studied problem in mechanism design without money. Procaccia and tennenholtz proposed and studied the problem back in 2009, where there are n agents on a line and the government will build a facility in a certain location given the agents reported information on their positions. Since every agent wants the facility to be closer to her, the government wants to make sure truth-telling is the best strategy for every agent while achieving some optimization objective. Since then, some bounds on the approximation ratios of the truthful mechanisms have been improved and new models are proposed. In this talk, we will briefly explain the story of the classic model and emphasize on therecent development on new models proposed by us and other groups. On the other hand, a resource auction model will also be briefly discussed where jobs bid for time slots from the virtual machines and the objective of the cloud is to elicit truth information from the jobs while maximizing the revenue.
学术报告二
报告题目：Service Matching with Uncertain Outcomes
报 告 人：齐向彤教授， 香港科技大学
报告时间： 2018年5月10日 3:00-4:00PM
报告地点： 广东工业大学工学一号馆216室
简介：
齐向彤博士是香港科技大学工业工程及决策分析学系教授。 他的主要研究兴趣是运筹学和博弈论，应用领域包括制造，物流，供应链，以及通讯网络等。他的学术期刊编辑经历包括IIE Transactions, IEEE TASE, APJOR 和 FSM，并为以下公司提供咨询服务：HIT, HACTL, Huawei等.
Dr. Xiangtong Qi is Professor in the Department of Industrial Engineering and Decision Analytics at The Hong Kong University of Science and Technology. His research interests are operations research and game theory, with applications in service, production, logistics, supply chain, and communication networks. His editorial experience includes Associate Editorship for IIE Transactions, IEEE TASE, APJOR and FSM. He has done industrial consulting projects for HIT, HACTL, Huawei, etc.
Abstract：We consider a model of assigning a set off servers to dynamically arrived demands, where the outcome is a random reward function of the demand and the number of servers assigned. We show that there exists one type of policy that is optimal to different reward functions. The model has potential for a variety of applications such as sharing economy, online ad auction, charity, and so on.
学术报告三
报告题目：Network Slicing: How Much Resource to Ask and How to Measure the Performance
报 告 人： 汪建平副教授， 香港城市大学
报告时间： 2018年5月10日 4:00-5:00PM
报告地点： 广东工业大学工学一号馆216室
简介：
汪建平博士是香港城市大学计算机系副教授。加入香港城市大学之前，她曾在美国南佐治亚大学，密西西比大学担任助理教授。她的主要研究兴趣是云计算，面向服务的网络，无线传感网络，光网络及网络性能测量。她是IEEE Transactions on Mobile Computing 以及 Optical Switching and Networking 杂志的编辑。她的研究获得美国自然科学基金，中国自然科学基金，香港研究局，香港创新科技局，谷歌，华为及美国Xinova公司资助。
Dr. Jianping Wang is an associate Professor in the department of Computer Science at City University of Hong Kong. Prior to joining City University of Hong Kong, she was an assistant professor at University of Mississippi and Georgia Southern University. Her main research interests are cloud computing, service oriented networking, wireless sensor networks, optical networking, and network performance measurements. She is an associate editor of IEEE Transactions on Mobile Computing and an associate editor of Optical Switching and Networking. Her research has been sponsored by US NSF, NSFC, Hong Kong Research Grant Council, Hong Kong Innovation Technology Council, Google, Huawei, and Xinova (US).
Abstract：In today's telecommunication industry, a grand challenge faced by many network service providers (NSPs) is how to cost-effectively allocate resources so as to agilely accommodate a variety of applications and services with diverse quality-of-service (QoS) requirements. To address this challenge, a promising technology is network slicing, with which tenants, can create different network slices with desired performance guarantees on the same network infrastructure. In the past three years, network slicing has attracted significant attention from academia and industry. However, most existing studies have focused on how to dynamically partition a physical network infrastructure into many network slices, and thus assumed that each tenant knows the right amount of resources which is necessary to accommodate certain traffic load with performance guarantees. This assumption causes a gap between tenants and NSPs, which compromises the adoption of network slicing. In this talk, we first introduce how to model the relationship among traffic load, resource, and performance. With this model, given the traffic load and desired performance, we can derive the amount of resource to be provisioned. We then introduce a network tomography based approach to infer a set of interested paths with the minimum number of monitors.
