Open Collections

Sadhukhan, Sayan

University of British Columbia

The vertical alignment optimization problem in road design seeks the optimal vertical alignment of a road at minimal cost, taking into account earthwork while meeting all safety and design requirements. This study introduces a unified model that enhances the quasi-network flow model by incorporating various material types, multiple hauling paths, and numerous roads within a network, as well as addressing the challenges associated with road side slopes. The model is devised to produce a vertical alignment that more accurately reflects the real-world scenarios of road construction, applying suitable constraints. Furthermore, we expand this multi-material, multi-haul, multi-road quasi-network flow model into a convex optimization framework, as opposed to a mixed-integer linear programming model, by redefining the volume constraints into a quadratic form. This model is capable of managing multiple material types while preserving the convexity of the model. We refer to our approach as the MRMH-QCQP-QNF model, and demonstrate its capability in approximating material volumes, particularly when benchmarked against the previous QCQP-QNF model. Additionally, our findings reveal that this model is adept at determining the optimal vertical alignment for larger road networks, a task at which the mixed-integer linear programming model falls short. Our convex model scales in polynomial time for larger road networks as compared to the exponential time complexity of the mixed integer linear programming model.

Text

2024-12-15

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/88120

Computer Science

University of British Columbia

UBCO

http://creativecommons.org/licenses/by-nc-nd/4.0/