Selecting the public transit projects with PCA-DP technique: The example of Xiamen City

• Combines principal component analysis (PCA) with the dynamic programming (DP) to prioritize transit projects.
• Integrates multi-criteria indicators into the dynamic decision process for multi-period project selection.
• Uses data from 15 cities of China to determine the appropriate transit projects for the empirical city.

Many cities in China are in the process of investing in public transit projects in order to reduce the growing problem of traffic congestion and enhance mobility of travelers. Cities all tend to pursue large public transit projects when it comes to investment options. There seems to be an urgent need to study these options from the standpoint of planning objectives and travel requirements. This paper compares the payoff of alternative projects of the public transit system for Xiamen City. We formulate the transit project selection problem as a mathematical model. With the proposed PCA-DP technique which integrates the principal component analysis (PCA) into a dynamic programming (DP) model, the optimal solution can be obtained based on indicators that cover four aspects of transit system performance evaluation: the level of service, the predicted relevant income and cost, and the expected external influence. The bi-hierarchy PCA model, constructed with data from 15 cities of China, generates the payoff function used as input to the DP model for the multi-phase transit project selection process for the City of Xiamen. The empirical analysis shows that for most of the scenarios, upgrading the existing bus rapid transit (BRT) into light rail transit (LRT) system and constructing a mass rail transit (MRT) routes in the next 10 years appear to be the most cost-effective option. Combined with other findings from this study, the result of our work can be used by practitioners to facilitate transit system planning and transportation project prioritization. The solution technique discussed provides a methodological framework that can be potentially used for the selection of a future transit system.