- We developed a three-stage approach to solve the problem efficiently: (i) create the possible sequences of bus driving assignments, called tours; (ii) create favorable tour patterns for different on-duty and off-duty arrangements; and (iii) choose the optimal selection of patterns to cover all driving assignments.
- We considered multiple objective function factors, union and regulatory rules and the implementability requirements at each stage.
- We worked in close collaboration with the driver planning group at Greyhound to ensure that tacit knowledge was included in the optimization model, the model solutions were implementable and the system was easy to configure.
- We illustrated results to end users with a business intelligence tool for detailed analysis and to agree on the final configuration.
- Our model demonstrated a 60% reduction in 6-day work weeks without compromising on driver wages, leading to higher driver satisfaction and retention.
- The model created more regular driver schedules, resulting in a 33% reduction in extra boards.
- Our model reduced layover and driver idle time costs by 5%.
- Planners were able to revamp driver plans within 15 minutes and then study the impacts, a dramatic improvement in the planning process.
- Rapid design and development in agile mode enabled the creation of implementable, optimized schedules within 8 months of beginning the project.
- The Driver Plan Optimization Group at Greyhound earned runner-up at FirstGroup’s 2018 Be First Awards for their significant contributions to making the company more efficient.