2019-12-27 | Junqi Hu: Scheduling in Queueing Systems with Specialized or Error-prone Servers
Cross-trained (flexible) servers are widely discussed as a useful tool to improve the performance of manufacturing and service systems. Our objective is to maximize the long-run average throughput of a multi-server queueing system via dynamic scheduling of flexible servers. In particular, we consider three cases which are practical in real life but seldom discussed in the literature: (1) a job might be broken and wasted when being processed by a server; (2) a job can be decomposed into multiple subtasks; (3) a job needs to be served by a group of servers as a team. We formulate the control problem as a Markov Decision Process (MDP). For smaller systems, we identify the optimal server allocation policy using MDP techniques. For larger systems, we obtain partial characterizations of the optimal policy through sample-path arguments and propose near-optimal heuristic policies.
Junqi Hu is a Ph.D. candidate in the H. Milton Stewart School of Industrial and Systems Engineering at the Georgia Institute of Technology. She received a B.A. in Economics and a B.S. in Mathematics from Wuhan University, China, in 2014, an M.S. in Operations Research from Georgia Tech in 2016. Her research interests are in dynamic control of manufacturing and service systems, Markov decision processes, and queueing networks.