The discrete space p-median model is the best known and most widely used location-allocation model. As data, it requires a vector of demands and a distance matrix. In real world applications both types of data are subject to estimation error. In this paper, simulation is used to investigate the relationships between input errors and the magnitude and range of output errors. The p-median model is quite robust in the face of both types of error, but tolerates demand error better than distance error.
