| We present three articles written to satisfy the requirements for the Ph.D. degree in Operations and Information Systems: The first is focused on the use of a bidimensional Markov model to compute the expected performance of an ambulance system using dynamic ambulance repositioning based on the use of a compliance table. The model is validated against a detailed discrete event simulation model, and we show that the ranking of the results obtained from multiple compliance tables is highly correlated with the ranking obtained from the discrete simulation model. The second paper deals with the problem of finding optimal or near-optimal compliance tables for an ambulance system. We propose a framework to classify optimization models and we use it to put the two models developed into context. The first model is an integer programming formulation that assumes the ambulances are always in compliance, while imposing constraints on the ambulance repositioning required. The second model takes the bidimensional Markov model and uses it in a heuristic search to find near-optimal compliance tables.;The final paper, although not directly related to ambulance operation, resulted from the implementation of a discrete event simulation of an ambulance system. In this paper we consider pre-computed routing information from any node to any other node in a road network. We compress it by taking advantage of the structural properties of the information and by transforming the problem into a traveling salesman problem which can be solved either to optimality via a solver or approximately via an insertion heuristic. |
