Stochastic optimization models for location and inventory prepositioning of disaster relief supplies

We consider the problem of preparing for a disaster season by determining where to open warehouses and how much relief item inventory preposition in each. Then, after each disaster, prepositioned items are distributed demand nodes during post-disaster phase, additional procured as needed. There is often uncertainty level, affected areas’ locations, items, usable fraction post-disaster, procurement quantity, arc capacity. To address uncertainty, we propose analyze two-stage stochastic programming (SP) distributionally robust optimization (DRO) models, assuming known unknown (ambiguous) distributions. The first second stages correspond pre- phases, respectively. also model that minimizes trade-off between considering distributional ambiguity following belief. obtain near-optimal solutions our SP using sample average approximation computationally efficient decomposition algorithm solve DRO models. conduct extensive experiments hurricane an earthquake case studies investigate these approaches computational operational performance.