Sustainable joint maintenance and renewal planning for urban railways via lexicographic optimization

The growing emphasis on sustainable transportation has intensified the need to reduce carbon emissions in railway maintenance and renewal activities. This paper proposes a lexicographic optimization framework for joint maintenance and renewal planning in urban railway networks, in which CO2 emissions are prioritized over economic cost. The problem is formulated as a mixed-integer linear program integrating key industrial constraints, including track possession windows, a degradation model, and activity grouping. A two-stage solution strategy is developed: the first stage identifies CO2-optimal solutions, while the second stage refines these solutions with respect to cost using either an exact MILP formulation or an iterated local search metaheuristic. Computational experiments on real-world network-level instance show that the proposed MILP–ILS approach preserves CO2 optimality while achieving notable cost reductions and significantly lower computation times compared to a fully exact lexicographic MILP. In several instances, the approach reduces network-level CO2 emissions by up to 50% while achieving cost savings of approximately 10% comparing to the company’s current strategy. Finally, an ɛ-constraint analysis is conducted to characterize the Pareto frontier between cost and CO2 emissions. The results confirm that substantial environmental gains can be achieved with moderate economic trade-offs, reinforcing the practical relevance of the proposed framework.