Modeling and algorithmic development of a staff scheduling problem

Abstract

Scheduling workers in a trucking system for stripping and loading trucks is a difficult and time consuming task that involves determining the optimal number of workers. Once the number has been determined the operations manager has to assign different trucks to the workers and has to determine the schedule of the workers. In the thesis,we develop a mathematical model to solve the scheduling problem, with attention focused on minimizing the number of workers required on a particular shift and on finding the feasible assignments of workers to trucks and vice-versa. We show that this is a set partitioning problem with an additional feasibility constraint which can be solved using the column generation technique. An inherent characteristic of this problem that makes it different from other set partitioning problems is that the arrival time and scheduled departure tin1e is different for different trucks. We illustrate the model and the proposed algorithmic approach by generating a schedule based on real data obtained from a trucking company.