Corridor location for developing new infrastructure such as transmission lines, roadways, and pipelines over terrain must consider numerous factors when determining the set of optimal route candidates. Previous researchers have cast this problem as a multiobjective least cost path problem, where competing objectives represent cost, environmental impact, and other major noncommensurate objectives. To fully characterize the optimal trade-off solution set, one must be able to generate both supported and unsupported nondominated solutions. Fortunately, the supported, nondominated solutions are relatively easy to identify by using a single-objective shortest path algorithm in conjunction with the noninferior set estimation method of Cohon, Church, and Sheer. But, finding the unsupported nondominated solutions can be nondeterministic polynomial time hard. This article proposes a heuristic approach that is capable of determining a Pareto frontier that is very near exact in polynomial time. This heuristic approach uses gateway node and gateway arc paths to generate a large set of spatially diverse locally optimal candidate solutions with few shortest path solver iterations, which are shown to represent a solution set that approximates to a high degree, the exact Pareto set. The method is described within the context of a bi-objective corridor location problem and applied to several data sets that have been used in past work. A comparison between this new approach and existing exact algorithms is provided. Overall, the new method is shown to be effective at finding a sizable number of the unsupported nondominated solutions in a very small amount of computational time.
