High-speed rail (HSR) systems planning and decision-making is a large-scale, multistage, intricate problem, where each stage represents a complex sub-problem. The solution approach to the multistage problem needs information exchange across the stages. A non-optimal planning and decision at any stage could yield a sub-optimal HSR system. This study proposes a novel multistage iterative HSR station location and alignment optimization framework (HSLAOF) to solve the main problem by decomposing it into the identification of (a) HSR potential regions and corridor, (b) station location, (c) shortest sequence/corridor of stations, and (d) collision-free alignment (horizontal and vertical) sub-problems. It integrates these sub-problems, which are solved separately using customized artificial intelligence- and motion planning-based metaheuristics such as particle swarm optimization for station location, ant colony optimization for shortest sequence of stations, path planner method for horizontal alignment and exploration, and exploitation-based ant colony optimization for vertical profile. In HSLAOF, the information from the upper and lower stages of the optimization process are exchanged and combined for a holistic optimized HSR system. Application of HSLAOF in the Mumbai-Ahmedabad HSR project yielded a solution that has 14.24 % lower total system cost, no environmental impact, 5.82 % better socio-economic impact, 21.14 % lesser noise and vibration impacted population, and 2.79 % higher station location utility than the one proposed by the experts. For predefined station location, the HSLAOF yielded alignment with 13.90 % lower total system cost and 24.00 % less noise and vibration impacted population than the one developed by experts. It required about 150 min to obtain a solution for each iteration.
Read full abstract