Emergency medical service location problem based on physical bounds using chance-constrained programming approach

Abstract

This paper studies a location and sizing problem for an emergency medical service system using distributionally robust chance-constrained programming approach. The medical demands are uncertain and described in an ambiguity set, which is constructed based on Wasserstein-metric. The location problem is modelled through minimising the total expected costs, including construction costs of facilities, purchase of ambulances and maintenance costs. The chance constraints are introduced to guarantee the service level of geographical areas. For the chance constraints, three reformulations such as CVaR-approximation, the mixed integer linear programme exact reformulation and the physically bounded bilinear reformulation are studied. A numerical study is conducted to illustrate the computational performance and objective values.