Campuses:

Absenteeism Prediction and Extra-board Driver Scheduling for Bus Transit Operations

Monday, December 3, 2018 - 3:00pm - 4:00pm
Lind 305
Qie He (University of Minnesota, Twin Cities)
In a tight labor market, it is essential for a public transport agency to manage its bus workforce efficiently to deliverer reliability services. In this talk, we present a data-driven decision tool to assist daily scheduling of extra-board drivers for Metro transit, the dominant public transportation service provider in the Twin cities. In practice, extra-board and overtime drivers are used to cover open work due to the absences of regular drivers. By analyzing the historical scheduling and driver absence data, we build a hierarchical logistic regression model to forecast daily driver absences. Building upon the prediction result, we develop a two-stage stochastic programming model to assist the dispatchers to determine optimal assignment of extra-board drivers and use of overtime. Our model's recommended policy reduces the expected total operating cost significantly while maintaining a low probability of losing service for certain routes.