专家讲座:Optimal Dynamic Assignment of a Flexible Server into Two Heterogeneous M/M/1 Queues in Non-Preemptive Service Discipline

发布者:周姝怡发布时间:2019-10-15浏览次数:2280


讲座题目:Optimal Dynamic Assignment of a Flexible Server into Two Heterogeneous M/M/1 Queues in Non-Preemptive Service Discipline

讲座专家:冯有翼教授(四川大学)

讲座时间:20191023日(周三)14:30

讲座地点:有线楼306会议室

讲座摘要:This paper studies effective server assignment for a queuing system of two parallel heterogeneous M/M/1 queues. Each queue is attended by a dedicate server with a specific service rate. A flexible server however can be assigned to either queue based on the congestion levels of both queues. Customers are served by a non-preemptive discipline. Such a model with heterogeneous dedicated and flexible servers is motivated by a practical problem of allocating hospital beds between two types of patients (called daily and selective patients) and has not been addressed in the queueing literature.  Our focus is on characterizing the structure of optimal assignment policies for the flexible server in the two-channel queuing systems. The analysis can be extended to similar queuing systems with multiple switchable servers. In addition, via the determination of the optimal assignment policy, we reveal that a modified mu c rule is still valid for the dynamic non-preemptive queuing setting.

讲座专家简介:

冯有翼教授现为四川大学商学院特聘教授。他曾是西班牙萨拉戈萨MIT全球物流教研中心绿色供应链终身讲座教授,也是MIT交通和物流中心的兼职研究员。曾任香港中文大学工学院系统工程与工程管理系副教授,香港城市大学商学院管理科学系访问教授,电子科技大学高端特聘教授,上海财经大学商学院特聘教授, INFORMS协会会员,全球供应链管理会议组委会成员等职。先后获得南京大学数学系数学学士、数学硕士,美国哥伦比亚大学工程与应用科学学院工业工程硕士、工业工程和运筹学博士学位。冯教授的研究方向包括:可持续发展的绿色(低碳)供应链管理,金融工程与风险管理,实物期权理论和应用,经济综合发展,能源合理利用,保护环境和控制污染的创新模型,能源需求和价格均衡预测,随机工业制造和服务系统控制与博弈,大规模零售和制造业供应链管理基础理论与优化算法。其研究成果广泛发表于《Management Science》、《Operations Research》、《IEEE Transactions on Automation Control》、《IIE Transactions》和《Naval Research Logistics Quarterly》等运营管理领域国际最顶尖的学术杂志。冯有翼教授还有丰富的产业咨询经验,先后担任过纽约Bearing证券、Marshall Lazer银行金融分析师和美国休斯顿安然公司天然气运输收益和风险研究项目经理。