Simultaneous optimization of 3D alignments and station locations for dedicated high‐speed railways. (22nd June 2021)

• Record Type: Journal Article
• Authors: Song, Taoran; Pu, Hao; Schonfeld, Paul; Zhang, Hong; Li, Wei; Hu, Jianping
• Is Part Of: Computer-aided civil and infrastructure engineering. Volume 37:Number 4(2022)
• Page Start: 405
• Page End: 426
• Publication Date: 2021-06-22
• Subjects: Civil engineering -- Data processing -- Periodicals; Computer-aided engineering -- Periodicals; 624.0285
• DOI: 10.1111/mice.12739
• Languages: English
• ISSNs: 1093-9687

Abstract

Abstract: Determinations of alignments and station locations are major tasks in railway design, which jointly and fundamentally influence the construction and operation of the whole project. Currently, although considerable efforts have been invested into the separate optimization of alignments or station locations, the complex interactions between alignments and stations, as well as their concurrent impacts on railways still need further investigations. Therefore, this study focuses on modeling and optimization of passenger railway alignments and station locations simultaneously. A novel optimization model is formulated for maximizing net present value (NPV) considering comprehensive costs and revenue from riders over the railway's design life. Specifically, the comprehensive cost includes construction and operation costs of alignments and stations. The railway ridership is quantified by combining travel time and trip coverage using a logit function. Then, the railway revenue is estimated and combined with costs through an economic growth model to obtain the NPV. To solve this model, two customized and mutually exclusive methods, that is, the station‐alignment‐integration (SAI) and alignment‐station‐integration (ASI) methods, are developed based on a particle swarm algorithm. The SAI primarily searches for possible stations in the landscape and then generates alignments to connect adjacent stations. The ASI directly yields the main alignment, on which potential stations are located by handling specific constraints. The model and methods are applied to an actual high‐speed railway. Results show that both SAI‐ and ASI‐generated railways have higher NPV values compared to the corresponding railway designed by experienced designers. Sensitivity analysis reveals that SAI is more flexible than ASI in solving actual cases. The convergence of the proposed methods is also discussed. With some changes in input parameters based on locally applicable design standards, the proposed methods are generally applicable to other cases or countries.