Optimized Rostering of Workforce Subject to Cyclic Requirements
Abstract
SNCF is a large railway transportation company that operates 365 days a year and 24 hours a day. In order to schedule a certain category of workers at train stations and ticket selling points, rosters are designed to cover a cyclical demand. However, the highly combinatorial nature of the rostering problem makes it very difficult to solve it manually, and experts spend a huge amount of time to make them legally feasible and to improve a certain number of preference criteria. This paper presents a mixed-integer programming model to address the cyclical rostering problem using patterns corresponding to feasible blocks of seven days and assigning them to each week of the roster. Some valid inequalities are presented to improve the linear relaxation of the model and thereby enhance computational performance. Implementation results are presented, including comparisons with an alternative daily-variables model.