Fast Approximate Algorithm for Robust Emergency System Design

This contribution is aimed at advanced rescue system designing. Since the modelled decisions are strategic and often have a long-term impact on accessibility and effectiveness of service providing, several aspects should be taken into account when a integer-programming model of the problem is created. One of them consists in system robustness that means its resistance to randomly occurring failures in the associated transportation network, which is used by ambulance vehicles when delivering rescue service to system users. To incorporate system robustness into the particular integer-programming model, a finite set of failure scenarios is usually formed. Then, the optimal robust rescue system design can be obtained by minimization of the highest detrimental impact of the individual scenarios. A simple rescue system design can be modelled as a weighted p-median problem. While a weighted p-median problem is easily solvable, the robust system design often causes computational difficulties due to the model size and its complicated structure. Therefore, the main goal of this contribution is to develop a fast algorithm for robust rescue system design, which enables to obtain a good approximate solution of the original problem in a short time. The theoretical explanation is here accompanied by a computational study.