Sensitivity analysis in continuous location models via interval analysis

Many continuous location models present the difficulty of the determination of some parameters which appear in the mathematical formulation of the problem, such as the weights associated with the destination points or the parameters of the distance functions. This uncertainty motivates the need to study the sensitivity of the model's solution to variations in the paramaters. In this paper, we consider a general continuous location model and assume that each parameter in the model is known to lie in a given interval. We address the problem of finding the set of optimal points to the problem when each parameter varies in its corresponding interval. To this end, we use ‘interval analysis’, a global optimization tool which can also be used to handle other optimization problems. An empirical study is carried out for a particular obnoxious facility location model to show the applicability of the method.