A Parallel Branch and Bound Algorithm for a Rolling Stock Cyclic Job Scheduling Problem to Minimize the Total Maintenance Cost

In this paper, we analyse an industry driven problem of a rolling stock cyclic job scheduling problem to minimize the total cost of preventive maintenance activities. We model it as a cyclic job scheduling problem on parallel machines with time and usage based deteriorating effects and maintenance to minimize the total maintenance total cost. To solve it, we propose an exact parallel branch and bound algorithm, which has a practical operational application to determine a monthly assignment of jobs to locomotives for small subsets of train routes and short optimization periods to gain considerable savings in a maintenance cost. Its time efficiency and range of applicability are verified during computational experiments.