A tailored branch-and-bound algorithm for routing a metro track inspection vehicle

Abstract

The track geometry data provide an important basis on which engineers can evaluate the status of a metro track. The manual routing process for inspection vehicles usually takes a long time, and the solution quality may be unsatisfactory. Considering both energy efficiency and the convenience of engineers, this article studies the routing problem for an inspection vehicle to collect the metro track geometry data with minimum total cost under complicated constraints according to industry practice. This work differs from previous studies in that the characteristics of the metro network and convenience for data-processing engineers are specifically considered. As this routing problem is an NP-hard problem, a novel graph transformation method and a tailored branch-and-bound algorithm with reference lower bounds are designed. Finally, a numerical case is studied in Beijing’s metro network. The numerical results show that the designed algorithm can solve the problem effectively, and it outperforms the genetic algorithm.