Considerable work has been done in optimizing highway alignments between two given endpoints. Previous attempts formulated the problem as a cost minimization problem and considered a number of user and operator costs. It would be desirable to consider the demand in the area served by the highway since demand can significantly influence the selection of a highway alignment. In this paper we formulate the highway optimization problem as a welfare maximization problem using relevant spatial information from a geographic information system (GIS). Genetic algorithms with eight problem-specific operators are used for optimal search. Unlike in traditional genetic algorithms the genes are encoded as real rather than binary numbers. We also develop a GIS-based algorithm to compute the price that users are willing to pay for service, and transmit it to genetic algorithms where operator costs are computed. An example of the proposed welfare maximization approach using real GIS information is presented. For the covering abstract see ITRD E128680.
Abstract