A three-stage stochastic facility routing model for disaster response planning

• A 3-stage stochastic model for facility location and distribution is proposed.
• Demand, vehicle availability and infrastructure state are stochastic parameters.
• Planning includes opening of facilities, allocation of supplies and distribution of aid.
• The model is evaluated on realistic test instances.
• Experiments show the value of using stochastic information to be significant.

This paper presents a three-stage mixed-integer stochastic programming model for disaster response planning, considering the opening of local distribution facilities, initial allocation of supplies, and last mile distribution of aid. The vehicles available for transportation, the state of the infrastructure and the demand of the potential beneficiaries are considered as stochastic elements. Extensive computational testing performed on realistic instances shows that the solutions produced by the stochastic programming model are significantly better than those produced by a deterministic expected value approach.