Integer Programming Approaches for Real-Time Traffic Management

Abstract

Traffic volumes are increasing in most of the major modes of transportation. As population centers grow in size and density, the increasing demand for public transit and delivery services leads to increased rail and road traffic. At the same time, it can be hard to expand the available infrastructure, as space is often at a premium where the population growth is fastest. Traffic management problems are a class of optimization problems where we seek to minimize delay or maximize throughput when using scarce infrastructure resources. In this thesis, I develop Integer Programming approaches suitable for practical, real-time application in air and rail traffic management. To achieve the needed efficiency, I introduce new modelling ideas and algorithmic developments. I make use of decomposition and iterative solution approaches. The results of this work include a working algorithm for optimal dispatching at Oslo Central Station.