A time-expanded network design model for staff allocation in mail centres

Abstract

We consider a staff allocation problem at a sequential sorting facility. In this facility, staff need to be assigned to work areas, through which commodities flow sequentially to be processed. Assigning staff optimally involves a trade-off between several different objectives, such as minimising the overall number of workers, as well as having fewer changes in the staff levels over time. While optimising for these, many operational requirements need to be met, including minimum processing volumes, correct ordering/processing of the commodities, and not exceeding staff resource constraints. We develop a deterministic time-expanded network flow model to solve the staff allocation problem. The model addresses the problem at a more granular timescale and with more operational constraints than previously used models. We use a lexicographical approach to deal with the multiple objectives. To demonstrate the model’s value, we apply it to a staff problem of a UK mail centre, showing that in the majority of cases, our model improves on current staffing practises on both objectives. We also show how the model performs in several different scenarios, including increasing total mail volumes and changing the proportion of letters and parcels to be sorted.