Cities are increasingly adopting bicycling as a cornerstone of sustainable transportation strategies, and e-bikes are contributing to this trend. This paper presents a systematic review of methods for selecting and prioritizing investments in cycling infrastructure, focusing on optimal Origin–Destination (O–D) pair identification and investment prioritization. The study highlights approaches that balance diverse objectives and stakeholder interests while examining the integration of machine learning techniques. Sixty papers were reviewed in accordance with PRISMA guidelines. The findings underscore the complexity of expanding urban cycling networks, as optimization involves balancing objectives such as maximizing ridership and minimizing travel effort, distance, or safety concerns. Achieving optimal or near-optimal solutions relies on methodological strategies ranging from GIS-based prioritization to weighted-sum formulations that balance competing planning criteria, to advanced optimization and heuristic techniques that explore large sets of candidate solutions. While methods such as Genetic Algorithms and Reinforcement Learning help manage problem size, there remains significant potential for applying Graph Representation Learning and clustering techniques to reduce computational complexity and extract O–D pairs from network and crowd-sourced data. Predicting the evolution of key indicators during plan implementation and assessing the impact of cycling networks on motorized traffic can provide valuable insights to decision-makers. This study provides a foundation for future research aimed at developing more effective and efficient frameworks for urban cycling network planning.
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号