Towards Improved Optimization Models for Locating Ambulances

by

Erhan Erkut
University of Alberta

The objective in many facility location problems is to minimize the total cost of transportation.  In contrast, the objective in selecting locations for ambulance stations is to maximize coverage, namely maximize the number of potential calls that can be reached within a specified time target.  Given the significance of this public decision problem, ambulance location problems have attracted the attention of operations researchers for at least 30 years.  In this talk we will focus on prescriptive ambulance location models.  We will emphasize two serious shortcomings of the traditional maximum covering model and introduce a new model that does not suffer from these shortcomings.  Our preliminary computational experience suggests that the new model is tractable and the solutions it offers are sufficiently different from other models presented in the literature to warrant its use in practice.  We will also discuss a number of new avenues of research to further increase realism in ambulance location problems.