Financing mode and scheme decision support for large urban rail transit projects: a revised case-based reasoning approach
Xian Zheng,Yiling Huang,Yan Liu,Zhong Zhang,Yongkui Li,Hang Yan
DOI: https://doi.org/10.1108/ecam-03-2023-0202
2024-10-25
Engineering Construction & Architectural Management
Abstract:Purpose As the complex influencing factors for financing decisions and limited information at the early project stage often render inappropriate financing mode and scheme (FMS) selection in the large-scale urban rail transit (URT) field, this study aims to identify the multiple influencing factors and establish a revised case-based reasoning (CBR) model by drawing on experience in historical URT projects to provide support for effective FMS decisions. Design/methodology/approach Our research proposes a two-phase, five-step CBR model for FMS decisions. We first establish a case database containing 116 large-scale URT projects and a multi-attribute FMS indicator system. Meanwhile, grey relational analysis (GRA), the entropy-revised G1 method and the time decay function have been employed to precisely revise the simple CBR model for selecting high-similarity cases. Then, the revised CBR model is verified by nine large-scale URT projects and a demonstration project to prove its decision accuracy and effectiveness. Findings We construct a similarity case indicator system of large-scale URT projects with 11 indicators across three attributes, in which local government fiscal pressure is considered the most influential indicator for FMS decision-making. Through the verification with typical URT projects, the accuracy of our revised CBR model can reach 89%. The identified high-similarity cases have been confirmed to be effective for recommending appropriate financing schemes matched with a specific financing mode. Originality/value This is the first study employing the CBR model, an artificial intelligence approach that simulates human cognition by learning from similar past experiences and cases to enhance the accuracy and reliability of FMS decisions. Based on the characteristics of the URT projects, we revise the CBR model in the case retrieval process to achieve a higher accuracy. The revised CBR model utilizes expert experience and historical information to provide a valuable auxiliary tool for guiding the relevant government departments in making systematic decisions at the early project stage with limited and ambiguous project information.
engineering, civil, industrial,management