UBC Undergraduate Research

Road design optimization with a surrogate function Pushak, Yasha


Planning new highways is a complex process requiring consideration of several non-trivial cost factors. An initial path is first refined by adjusting its vertical alignment to reduce earthwork cutting, filing, and hauling costs. This optimization problem is then used as a sub-problem that needs to be repeatedly solved with different horizontal alignments, to find an optimal path. The complete optimization process requires significant computation time to solve real-world problems. A multi objective surrogate function, which approximates the earthwork and paving costs and can be quickly evaluated, can be used to reduce the running time of the full optimization algorithm. This thesis builds upon an existing surrogate function, using a \Mass Haul" to approximate the earthwork hauling costs. Other constraints made to simplify the former surrogate are relaxed to improve the quality of the solutions returned by the surrogate. Numerical results compare solutions returned by the two versions of the surrogate and investigate the Pareto front between the earthwork and paving costs.

Item Citations and Data


Attribution-NonCommercial-NoDerivs 2.5 Canada