University of California Transportation Center

Drivers account for up to 80% of the operational cost of transit agencies. This dissertation provides a method for improving the productivity of this workforce by introducing flexible contracts. Under these contracts drivers do not work the same number of hours every day. However, the number days and hours worked every week are kept constant. These contracts allow the agency to find a better match between the number of drivers needed and hired. Since people's preferences vary, introducing this flexibility should benefit both the agency and the drivers.

Driver contracts are offered simultaneously to all drivers. Because drivers may be absent to work unexpectedly, these contracts must be designed to overstaff the system. However, current methodologies to determine these contracts fail to incorporate this uncertainty. There is also a lack of understanding as to how parameters of the problem such as driver wage, and the likelihood of absenteeism will impact the total cost. This dissertation aims to fill this gap. The methodology introduced here considers absenteeism, and provides cost sensitivities.

Further cost savings can be achieved if these contracts are synchronized with the trips operated by the agency. To achieve this goal, this dissertation extends the contract scheduling work and investigates the opposite problem where contracts are given and timetables must be identified. Two different methodologies are used to study this timetable design problem: continuum approximation and numerical optimization. The continuum approximation work incorporates uncertainties and provides cost sensitivities. The numerical optimization approach is used to validate the approximated results. In the latter work cuts are designed to solve an otherwise intractable problem. These cuts are also valid for the production lot-sizing problem.