Pre-disaster investment decisions for strengthening a highway network

transportation network fortifying
Peeta et al. \cite{PeetaEtal2010} addressed a pre-disaster planning of stochastic transportation network retrofitting problem for earthquake disaster response. Having considering links failure independent,the authors define only one scenario (survival probability) for each link, and also assume that link investment decisions are binary integers. Consequently,two-stage stochastic program is presented and seeks to identify subset of links to invest under available budget in the first stage, and determine minimum response time of the network in second stage. For solving the problem, the integrality of investmetn variables was relaxed and Taylor series expansion applied to objective function and a monotonic non-increasing multilinear function is obtained. Using first order terms of the function results a knapsack problem whose optimal solution is shown to be a local optimum to the original problem. Computational experiment is carried out on Istanbul highway network with up to 25 nodes and 30 links.