Title: Stochastic network design and its applications in transportation and energy infrastructure system planning

Speaker: Yueyue Fan, University of California, Davis

Date/Time: Monday, April 26, 2010 at 9:00 – 10:00am       

Location: CSRI Building/Room 90 (Sandia NM)

Brief Abstract: In this talk, a general two-stage stochastic programming network design model will be presented with discussion on possible means of overcoming some of the modeling and computational challenges.  The applicability of this model in infrastructure system planning and management will be demonstrated via two examples.  The first application area is in transportation system disaster mitigation.  Given limited resources, the goal is to choose the best subset of network components to protect while allowing the network users to follow their own best-perceived routing decisions. This problem is formulated as a two-stage stochastic programming problem with equilibrium constraints, where the objective is to minimize the total expected physical and social loss caused by potential disasters.  The second application area is in renewable energy supply chain optimization.  A two-stage stochastic supply chain model is established to find the most cost-effective system design and energy feedstock recourse allocation strategies under uncertainties.  This mathematical model is used to explore the economic viability of waste-based biofuel production in California and the infrastructure requirements to support such a system.  In both examples, decomposition-based solution algorithms are implemented to overcome the numerical challenges caused by the large sizes of the problems.  The methodologies discussed in this seminar are relevant to other types of networks such as information, biology, and social networks. 

CSRI POC: Jean-Paul Watson, (505) 845-8887