Stochastic Dynamic Programming Solution of a Risk-Adjusted Disaster Preparedness and Relief Distribution Problem

This chapter proposes a multistage stochastic optimization framework that dynamically updates the purchasing and distribution decisions of emergency commodities in the aftermath of an earthquake. Furthermore, the models consider the risk of exceeding the budget levels at any stage through chance constraints, which are then converted to Conditional Value-at-Risk constraints. Compared to the previous papers, our framework provides the flexibility of adjusting the level of conservativeness to the users by changing risk related parameters. Under some conditions, the resulting linear programming problems are solved through the Stochastic Dual Dynamic Programming algorithm. The preliminary numerical results are encouraging.