Models and solution approaches for facility location of medical supplies for large-scale emergencies.

摘要

Research on facility location in a context of emergency services has focused on regular small-scale emergencies such as household fires or vehicle accidents, in which only a small amount of demand for medical service occurs. Most facility location models in the literature, therefore, consider providing a single facility (possibly including a few backup facilities) to cover a demand point. Such emergency models however typically have not considered conditions of large-scale emergencies, where a tremendous demand and low frequency combine to create situations with insufficient supplies and large uncertain demands. These conditions require a modification in the definition of facility coverage to allow for redundant facility placements and tiered facility services to ensure an acceptable form of coverage of all demand areas.;The purpose of this research is to develop models and solution approaches that are suitable to the facility location of medical supplies for large-scale emergencies. We formulate different integer models with objective functions to maximize the population coverage, or minimize the average/maximal distance from the facilities to the demand points. Since there can be different scenarios during emergencies, we also propose a regret model to find a globally optimal solution across scenarios. These models are computationally intractable, and therefore we develop heuristics to solve the problems. We perform experiments to evaluate our models and heuristics based on illustrative large-scale emergencies.