Decision support for siting problems

Despite the development of mathematics of location theory and despite its obvious economic impact, few applications have been developed and are actually in use to support decision makers in siting decisions. The obstacles that hinder a more widespread exploitation of mathematical results are twofold: the intrinsic difficulties of the relevant problems and the tradeoff to be balanced between the different objectives. The article presents the results obtained in the implementation of a Decision Support system applied to a specific siting problem: the location of installations for industrial waste management. This DSS is based on multicriteria decision analysis for the best siting of plants, optimizing costs and environmental impacts. The proposed approach individuates a hierarchy of objectives, where at top-level a k-best solution of the problem of minimizing a distance measure is computed, to reduce the space of alternatives. Specifically, the problem solved at top level has been mathematically formulated as a 0/1 fixed cost transportation problem (FCTP). This is an NP-hard combinatorial optimization problem that can only be solved heuristically for real world problem sizes. We propose a heuristic procedure, nearly optimal for the data sets we manage. The k-best solutions of the single objective problem are then combined to produce efficient alternatives to be further evaluated by means of multicriteria methods. Computational results both of the FCTP and of the whole systems are provided.