View article

This article describes several heuristics for the construction of a rapid transit alignment. The objective is the maximization of the total origin-destination demand covered by the alignment. Computational results show that the best results are provided by a simple greedy extension heuristic. This conclusion is confirmed on the Sevilla data for scenarios when the upper bound for inter-station distance is greater than 1250 m. Otherwise, when those upper bounds are smaller (750 m and 1000 m), an insertion heuristic followed by a post-optimization phase yields the best results. Computational times are always insignificant.