Z. Su, A. Núñez, A. Jamshidi, S. Baldi, Z. Li, R. Dollevoet, and B. De Schutter, "Model predictive control for maintenance operations planning of railway infrastructures," in Computational Logistics (Proceedings of the 6th International Conference on Computational Logistics (ICCL'15), Delft, The Netherlands, Sept. 2015) (F. Corman, S. Voß, and R.R. Negenborn, eds.), Lecture Notes in Computer Science, Cham, Switzerland: Springer, ISBN 978-3-319-24263-7, pp. 673-688, 2015.
This paper develops a new decision making method for optimal planning of railway maintenance operations using hybrid Model Predictive Control (MPC). A linear dynamic model is used to describe the evolution of the health condition of a segment of the railway track. The hybrid characteristics arise from the three possible control actions: performing no maintenance, performing corrective maintenance, or doing a replacement. A detailed procedure for transforming the linear system with switched input, and recasting the transformed problem into a standard mixed integer quadratic programming problem is presented. The merits of the proposed MPC approach for designing railway track maintenance plans are demonstrated using a case study with numerical simulations. The results highlight the potential of MPC to improve condition-based maintenance procedures for railway infrastructure.