This paper addresses the rota scheduling problem (RSP) which aims to assign a set of working days, free days and reserve shifts to driver groups such that management considerations, labor laws and the preferences of drivers are considered. Currently, the reserve shifts in the RSP are evenly planned for all drivers (for example: a certain percentage is planned for reserve shifts) without considering more detailed information such as historical or weekday-depended sickness absence rates. The current practice in the case that the absence rate exceeds the available reserve personnel is that additional drivers are called manually on demand. This not only causes discontent for the drivers, but also creates a day-to-day managerial burden for the bus company.
In order to reduce the discrepancy between a planned roster and the actual one, a new stochastic optimization model for the RSP is formulated and is compared to the deterministic optimization model. The scenarios of the stochastic problem are created depending on historical and weekday-dependent absence rates. In addition to the present reserve optional attendance shifts are introduced. This offers better conditions for the reserve drivers as well as enables a specific reaction on the sickness absence rate of the day by using optional reserves.