Two-stage likelihood robust linear program with application to water allocation under uncertainty

We adapt and extend the likelihood robust optimization method recently proposed by Wang, Glynn, and Ye for the newsvendor problem to a more general two-stage setting. We examine the value of collecting additional data and the cost of finding a solution robust to an ambiguous probability distribution. A decomposition-based solution algorithm to solve the resulting model is given. We apply the model to examine a long-term water allocation problem in the southeast area of Tucson, AZ under ambiguous distribution of future available supply and demand and present computational results.