Scenario planning approaches to facility location: Models and solution methods

In this dissertation we consider new models and solution methodologies for strategic facility location. Specifically, we develop scenario planning-based models which incorporate the uncertainty inherent in long term planning. The models also allow decision makers to incorporate a measure of risk aversion, thus deriving solutions which are better tailored to his/her business environment.;We first explore the alpha-Reliable P-Minimax Regret problem, a generalization of both the P-median problem (Hakimi 1964) and the minimax regret objective (Savage 1951). Properties of model solutions are explored and computational results are presented. The explosive size of this formulation is characteristic of scenario planning models and leads to computational intractability. We thus consider changes in the model to improve computational performance and finally develop a metaheuristic based on evolutionary programming. We also extend the metaheuristic to estimate the efficient frontier of solutions for a multi-objective version of the problem.;We next develop the Demand Node Based Satisfaction problem. This model is formulated first with a maximal covering-type objective (Church and ReVelle 1974) and then as a set covering problem (Toregas et al. 1971). Problem analysis and computational performance are presented. The maximal covering formulation presents a significant computational hurdle, and thus heuristics are again employed to make solutions tractable.;Finally, we introduce the Scenario-Based Satisfaction problem and discuss additional formulations which extend from the research presented in this thesis.