Kernel-like Search for Robust Emergency System Designing

Abstract

Emergency service system, which satisfies randomly emerging demands of public for necessary treatment, is determined by deployment of limited number of service centers at positions from a given set of possible locations. The objective is to minimize average response time of the nearest ambulance vehicle usually located at a service center. The robust service system is designed to comply with specified scenarios by minimizing the maximal value of the above mentioned objective functions corresponding to the particular scenarios, which represent consequences of random failures in the road network. The detrimental events may correspond to congestion, disruptions or blockages of roads. The robust emergency system design problem can be modeled by means of mathematical programming. The model includes scenarios and the associated link-up constraints, which connect average response time connected with individual scenarios to the general objective function, which is maximum of these objective functions. The min-max link-up constraints and the cardinality of the scenario set represent an undesirable burden in any solving process used for design solution. Within this paper, we present a kernel-like search algorithm, which tries to replace the solving process of the huge problem above by a series of smaller problems, which deal with either small subset of scenarios or reduced set of possible center locations.