A multi-objective grouping Harmony Search algorithm for the optimal distribution of 24-hour medical emergency units

Abstract

This paper presents a novel multi-objective heuristic approach for the efficient distribution of 24-h emergency units. This paradigm is essentially a facility location problem that involves determining the optimum locations, within the existing health care centers, where to deploy 24-h emergency resources, as well as an efficient assignment of patients to such newly placed resources through the existing medical care infrastructure. The formulation of the underlying NP-complete problem is based on a bi-objective distance and cost metric, which is tackled in our approach by combining a Harmony Search algorithm with a grouping encoding and a non-dominated solution sorting strategy. Additionally, the nominal grouping encoding procedure has been redefined in order to reduce the dimension of the search space, thus allowing for a higher efficiency of the searching process. Extensive simulations in a real scenario – based on the geographic location of medical centers over the provinces of Guadalajara and Cuenca (Spain) – show that the proposed algorithm is statistically robust and provides a wide range of feasible solutions, hence offering multiple alternatives for the distribution of emergency units.