Optimal road facility spare parts location with continuum approximation

Abstract

This study presents a new methodology for optimizing the location of spare parts depots for expressway facilities using a continuum approximation (CA) approach. The increasing importance of expressway facility asset management necessitates efficient strategies for minimizing both user impact during failures and the costs of maintaining spare parts depots. In this study, we first formulate a model within the framework of dynamic facility location planning (DFLP), a type of integer programming (IP), to optimize the spare parts location plan, taking into account the failure processes of expressway facilities. Traditional IP models are computationally intensive when applied to large-scale networks. To address this, we adapt the CA approach, traditionally used for facility location problems in Euclidean spaces, to handle network distances by embedding the road network into a new Euclidean space using the Isomap algorithm. The proposed methodology was then applied to spare parts location optimization problems of electronic toll collection (ETC) systems in a real-world expressway network in Japan. The results demonstrate that the proposed methodology significantly reduced the optimization computation time by 85.6% to 97.9% compared with an existing method, showcasing a substantial improvement in computational efficiency while also obtaining near-optimal solutions.