TITLE: A Stochastic optimal control problem in workforce management

SPEAKER: Xuefeng Gao

ABSTRACT:

An important problem in workforce management involves determining a company's staffing portfolio from a mix of different sources. We consider the special case of determining a company's positions in its internal and external workforce given the uncertainty and volatility of growing/emerging demand, and the risks associated with these positions. To solve this problem, we take a risk hedging approach and formulate a stochastic optimal control problem for maximizing the expected discounted profit over an infinite horizon. Our main results include deriving the optimal workforce management policy and establishing that it is of base stock type (2-sided). Numerical experiments have been conducted to provide insights and understand the optimal policy. Both our theoretical results and numerical experiments demonstrate that dynamic alignment of workforce in response to changing/volatile demand is critical. This is joint work with Yingdong Lu, Mayank Sharma and Mark Squillante at IBM Research.

Bio:

Xuefeng Gao is currently a Ph.D. student in the Department of Industrial and Systems Engineering at Georgia Tech. His advisors are Professor Jim Dai and Professor Ton Dieker. He received B.S. in Mathematics from Peking University, Beijing in 2008. He is interested in Applied Probability and Stochastic Processes, Stochastic Optimization and Market Microstructure.