The p-median problems deal with decisions of locating p facilities (medians) in a network, minimizing the sum of all distances from each vertex to its nearest facility. If demand information is available for each vertex of the network, then values on the capacity of each facility may be present (capacitated p-median problems). Facility location models have been proposed as decision making tools, mainly when geographic information systems (GIS) can be used to capture, store and analyze the data of the problems. In this work we present the integration of a p-median algorithm to both ArcView, a GIS by ESRI (Environmental Systems Research Institute, Inc.), and SPRING, a GIS developed by INPE (National Institute for Space Research). The computer programme that has been integrated to these geographic information systems implements a recent approach of Lagrangean/surrogate heuristic which uses location-allocation heuristics in order to search for the prime feasibility of intermediate dual solutions. The paper presents some computational tests which have been conducted with real data from the city of São José dos Campos, representing problems with up to 3280 vertices and 1141 medians, for the uncapacitated problem.
Facility Location; Geographical Information Systems; Lagrangean Heuristics
