Robust Resource Allocation
by CHEN Wenqing
Speaker: |
|
Date:
|
03 October 2008 (Friday)
|
|
|
Abstract
We are interested in resource allocation problems in the context of disaster relief. Since disasters occur only occasionally, it is not realistic to maintain a high level stock at all cities. It is a challenging problem to decide how to achieve a high probability that all cities' demand could be met if a disaster occurs, while maintaining a lower stock cost. We formulate such a resource allocation problem as a chance constrained model. We propose different tractable approximations of individual chance constraint models using robust optimization on a variety of uncertainty sets and show their interesting connections with bounds on the condition-value-at-risk (CVaR) measure. Upon that, a new approach for approximating joint chance constrained models is developed. The new approach builds on a classical worst case bound for order statistics problem, and is applicable even if the constraints are correlated. Computationally, the new approach converts a joint chance constraint to a set of second-order-cone constraints; therefore the resulting optimization model is more tractable. The new approach is applied to the network resource allocation problem with very successful computational results, which improve significantly compared to the current standard approach.
About the speaker
Dr. Chen Wenqing is a senior research analyst in Singapore Land Transport Authority. Before joining LTA, she had 2 year experience in ILOG Asia Pacific as Technical Account Manager and 5 year experience in China Eastern Airlines as Air Traffic Manager. She was involved in the design of Singapore land transport data warehouse and Shanghai travel agency information system. Besides, she had consulting experiences with front-end semiconductor production planning, power generation scheduling, energy market pricing and so on.
She received Ph.D. from Decision Science Department, National University of Singapore in 2007. She also holds dual Bachelor degrees in Engineering Economics and Computer Science from Shanghai Jiaotong University. Her research interest lies in robust optimization, transportation planning, and inventory planning.