| In this thesis we study the impact of facility failure on strategic facility location problems. In particular, we study a classical problem in Location Theory: the p-Median problem on a network. As is commonly done when p-Median problems are investigated, we also discuss the impact on the p-Center problem, another classical problem in the literature. The classical approach considers that facilities are reliable and service is always available whenever a customer needs it.; Therefore, we start by considering that facilities may fail. This consideration has strong ties with current events in North America where terrorist acts have forced governments, in particular the Government of the United States, plan for planning service disruption. It is crucial to find ways to minimize the impact of disruptions caused by criminal and terrorist acts. Although this is currently an important issue, the models provided in this thesis address the problems caused in any type of disruption, whether caused by criminal acts or arising from the natural failure which all systems may undergo sometime. In this paper it is considered that disruptions will eventually occur. We propose that by carefully designing a strategic facility location plan, decision makers can minimize the long term effects of disruptions.; Although the major component of this study is the variations on the p-Median problem, we also include a model that minimizes the expected maximum distance traveled by each customer subjected to some reliability constraint. The intention is to stimulate future research considering facilities' failure in a mini-max setting.; The results of this work includes exact solutions for special cases, heuristic solutions to complexes, guaranteed worst case bound for some heuristics and insights into the solution behavior as a function of the probability of the facility disruption vector. |
