Tuesday, March 13 at 4:15pm
Frank H.T. Rhodes Hall, 253
This talk is concerned with workforce management, a component of business operations that involves employing the available workforce as effectively as possible. In particular, our objective is to understand how to assign workers to tasks in real time as a function of system state with the goal of achieving the best possible production rate. Our approach involves modeling the business under investigation as a queueing network and the workers as servers. Jobs can differ in terms of both the magnitude and nature of the required work and workers can have different abilities with respect to what tasks they can handle and how efficiently they can accomplish these tasks both individually and in teams. Under these conditions, our objective is to determine optimal dynamic worker assignment policies that take into account the abilities of the respective workers (on their own and in various teams). Most of the talk will focus on Markovian tandem queues but we will also provide results for more general systems under more restrictive assumptions.
This is joint work with Sigrun Andradottir and Douglas G. Down.