Finding effective strategies of allocating limited mitigation resources to critical infrastructure system components for protection, response, and recovery is among the central tasks of disaster mitigation and management.
Dr. Liu’s dissertation tackles the pre-disaster network protection problem, a specific instance of the above general resource allocation problem, of determining which network components should be protected (e.g. retrofitted or strengthened) before disasters given resource constraints. This problem is not well solved in practice primarily due to the challenge of making protection decisions under uncertainty since disasters are not realized yet at the time of decision making and hence uncertain to decision makers.
Dr. Liu’s dissertation develops a rigorous approach based on stochastic programming and network optimization with the capability of capturing system component interdependency and explicitly incorporating uncertainty. It formulates the problem as a two-stage risk averse stochastic program which seeks a balance between minimizing expected system cost and reducing system cost variation. Two efficient numerical algorithms are developed to solve the two variants of the problem. Numerical experiment results demonstrate the superiority of the stochastic programming approach to deterministic methods (e.g. scenario analysis).
Dr. Liu is currently a postdoctoral researcher at Oak Ridge National Laboratory.
For the complete publication, view here