On finding dissimilar paths

Given a transportation network, this paper considers the problem of finding a number of spatially dissimilar paths between an origin and a destination. A number of dissimilar paths can be useful in solving capacitated flow problems or in selecting routes for hazardous materials. A critical discussion of three existing methods for the generation of spatially dissimilar paths is offered, and computational experience using these methods is reported. As an alternative method, the generation of a large set of candidate paths, and the selection of a subset using a dispersion model which maximizes the minimum dissimilarity in the selected subset is proposed. Computational results with this method are encouraging.