Incontrol recently finished a project for the Dutch Driving Test Organization, CBR. CBR assesses driving licenses in the first place and medical ability of drivers as well as professional skills of employees working in the field of transportation and logistics in the second place. Most important, CBR offers a wide range of exams for passenger cars, trucks, motor cycles and cars with trailers. To this end, CBR needs to carefully determine the number of exam slots and examiners. Exams, in this article, are referred to as the real driving tests instead of the theoretical exams.

Accordingly, CBR consulted Incontrol with the following short and long term objectives:

1. Insights in and objective quantitative support fir the "real" waiting times that candidate's experience.
2. Suggestions for and quantitative results on the effects of adaptations for the reservation system.
3. A user-friendly simulation application for capacity studies.

Process and Simulation Modeling

The project only focused on products for passenger cars: exams it self, possible re-sits and an interim test (a preparation for the exam). CBR currently realizes roughly 400.000 exams, including re-sits, on a yearly basis.

The business process of CBR is characterized by the following components:

• Arrival of candidates at driving schools
• Durations for candidates at driving schools (exams, re-exams, interim exams)
• Capacities for exams, re-sits and interim tests
• Release policy of capacity
• Purchase behavior of driving schools

CBR releases 50% of the available capacity at once 26 weeks beforehand. The other 50% is released in even portions from 25 weeks until 5 weeks beforehand. Driving schools make a prognosis of expected need for exams based upon the arrival pattern of candidates and expected durations for lessoning. Based upon this prognosis the driving schools should then purchase their expected demands from the released capacity as good as possible. The release policy and the purchase behavior are called the reservation system. Waiting times arise where actual demand in certain weeks exceeds the available capacity. The reservation time, mentioned before, is defined as the period up to the first possible week from the current moment to purchase an exam.

After the modeling and development phase of the simulation model, experiments were conducted to quantify the performance of CBR in terms of waiting and reservation times.

Experiments and analysis

First of all, CBR aimed to obtain insight in pure waiting times for candidates having to wait for their exam when ready with lessoning. There existed a kind of misunderstanding between driving schools and CBR about waiting and reservation times. These reservations times don't say much about the actual waiting times that candidates experience. A base model was created to compute the pure waiting times for candidates omitting the release policy and the purchase behavior. These two aspects distort the analysis of pure waiting times.

Thereafter the basic model was extended to include the release policy of exam capacity by CBR as well as the purchase behavior of driving schools. A distinction had to be made in large and small (one person) driving schools. A large driving school generally purchases far ahead, a small one just a couple of weeks.

By experiments the following main conclusions could be drawn:

1. The pure waiting times for candidates when ready with lessoning are most reasonable (approximately 95% no more than 1 week and almost always within 3 weeks)
2. The available capacity could meet the actual demand pretty well in spite of the variability's in the process (arrival of candidates, lessoning periods and chances of passing exams)
3. The long reservation times that CBR was confronted with are largely due to the reservation system rather than a structural shortage of capacity from CBR. The reservation system did not influence the overall waiting time performance. However, large driving schools took advantage of their wider purchase potential while candidates of small driving schools experienced seriously long waiting times.

Improvements

Based upon these experiments, CBR proposed system improvements to the associations for driving schools which are beneficial for all three parties involved. Driving schools experience a more transparent reservation system where they get the guarantee to reserve and purchase capacity 6 till 7 weeks beforehand. This way, the reservation competition between driving schools is eliminated. Also, candidates are guaranteed short waiting times when ready for exam after lessoning. At the last place, for CBR these improvements will result in a clearer view on actual capacity demand which makes the planning of capacity easier.

Future plans

CBR is at the point of closing agreements with the driving school associations on the proposed improvement of the reservation system and it's implementation. It also intends to reshape the simulation model into a user friendly application. This simulation application can then be used by planners of CBR to perform waiting time analysis and capacity studies on a regular basis for different regions and for different periods.