Optimal Multi-Commodity Network Flow of Autonomous Vehicles in Urban Traffic

Abstract: The focus of this thesis was to find, visualize and analyze the optimal flow of autonomous vehicles in urban traffic with respect to fuel consumption using linear optimization. Two different formulations based on multi-commodity network flow were implemented which resulted in a static and a dynamic model of the traffic. The static model was applied to Kungsholmen, an urban district in central Stockholm, Sweden, while the dynamic model was considered on a small-scale. These models led to large linear programs which were solved by applying different algorithms to the problems in various software. The results suggested that the developed models were adequate approximations of the urban traffic flow for the chosen parameter set-ups. It was concluded that the traffic in a system of autonomous vehicles may be optimally planned using similar models as formulated in this study.