Strategies to Enhance the Performance of Path-Based Static Traffic Assignment Algorithms

There are three important stages of path-based algorithms (PBAs) for solving the static user equilibrium traffic assignment problem (STA): finding shortest paths between various origins and destinations based on the present flow conditions to update the path set, updating path flows based on the move direction of the PBA, and updating the link flows and costs. This article proposes strategies to improve the computational efficiency of these three stages. The first strategy provides a simple method to preclude the through-routing via the zone centroid and helps to avoid unrealistic flow without affecting the flow update process of a PBA. The second strategy seeks to improve the efficiency of the path flow update process by circumventing unnecessary computation. The third strategy proposes faster link flow and link cost update processes along with a link data structure to support it. The computational experiments using two recently developed PBAs validate the effectiveness of these strategies and help to understand their rationale. The strategies are significant from both theoretical and practical perspectives. From a theoretical viewpoint, they help in designing an efficient execution process for PBAs and provide an improved common platform for comparing their performances. For practice, they can reduce the computational cost in finding the solution of the STA without increasing the complexity of the execution of the algorithm.