Decision support algorithms for short-term preventive maintenance planning of rolling stock

We consider the problem of optimizing rolling stock operations, including preventive maintenance, to enhance technical availability. Our approach accounts for various types of rolling stock and their multi-level maintenance schedules, integrating both time-based and distance-based cycles, a combination seldom explored in existing literature. Given the complexity of this problem, we focus on short-term planning, as it provides decision makers with implications for practice regarding current operational challenges and may also influence strategic decisions. We propose a branch and bound approach to derive optimal solutions for rolling stock operations, addressing limitations in traditional manual planning commonly used in business practice. Additionally, we developed an efficient heuristic to find a satisfactory solution in a short time. Simulations for real data show that our algorithms can provide feasible rolling stock schedules that significantly improve fleet availability in the enterprise. This improvement not only reduces maintenance and downtime costs but also enables the fulfillment of more transport orders, aligning with business priorities. (original abstract)