On finding dissimilar Pareto-optimal paths

The aim of the present paper is to provide a methodology for finding a set of alternative paths between an origin and a destination site on which routing one or a set of dangerous goods. Finding a set of paths allows one to equally distribute the total risk among the population exposed. The concept of equity of risk is here related to the concept of determining spatially dissimilar paths. We divide our approach into two phases. In the first phase we find a set of Pareto-Optimal paths between an origin and a destination, by implementing a multicriteria shortest path algorithm. In the second one, for each path previously found, and by using a geographical information system, we construct a Buffer Zone approximating the impact area of a material being released after an accident. Based on these Buffer Zones, a dissimilarity index between every pair of paths can be derived in order to find the most spatially different routes. We then compare our method with an iterative penalty method and discuss computational results based both on a real application and on test problems.