About Us 
Careers Students
Traffic congestion negatively impacts both residents and economic activities in affected areas. Since the costs of congestion are very high, numerous transportation research projects have been commissioned to find appropriate policies or measures to tackle these problems. This thesis is designed to analyze and model the way that road travellers react to these congestion relief policies and so provide pointers as to which projects and policies should be further developed.
Approach
A family of traffic assignment models is often used to analyze the effects of traffic congestion relief measures. In order to find the most appropriate model within this family, this thesis reviewed different models. Firstly, our research demonstrated that traffic assignment models can be classified into several different types: There are time-dependent and time-independent models and these can be further differentiated on the basis of stochastic and deterministic traveller knowledge (whether or not travellers’ perception errors are taken into account) and stochastic and deterministic networks (whether or not network uncertainty is taken into account). This thesis described two types of models in detail: (i) Time-independent models with deterministic traveller knowledge and deterministic networks — called the Deterministic Network Deterministic User Equilibrium (DN-DUE) and (ii) time-independent models with stochastic traveller knowledge and deterministic networks — known as the Deterministic Network Stochastic User Equilibrium (DN-SUE). Within this thesis, the DN-DUE model was applied to two sample networks. The performance of different traffic assignment models in analyzing congestion relief measures was also discussed.
Results
The use of traffic assignment models in analyzing congestion relief measures is promising: With the exception of a few measures, these models perform very well. From this thesis, we have concluded that the DN-SUE is the most appropriate model for analyzing congestion relief measures, because it maintains a good balance between reality and task difficulty. The Frank-Wolfe algorithm can be applied to this model, which is simple to implement and supplies good results. Further research is necessary when applying this algorithm to larger networks, since computation time then becomes a problem
Careers Students
Traffic Assignment
Traffic congestion negatively impacts both residents and economic activities in affected areas. Since the costs of congestion are very high, numerous transportation research projects have been commissioned to find appropriate policies or measures to tackle these problems. This thesis is designed to analyze and model the way that road travellers react to these congestion relief policies and so provide pointers as to which projects and policies should be further developed.
Approach
A family of traffic assignment models is often used to analyze the effects of traffic congestion relief measures. In order to find the most appropriate model within this family, this thesis reviewed different models. Firstly, our research demonstrated that traffic assignment models can be classified into several different types: There are time-dependent and time-independent models and these can be further differentiated on the basis of stochastic and deterministic traveller knowledge (whether or not travellers’ perception errors are taken into account) and stochastic and deterministic networks (whether or not network uncertainty is taken into account). This thesis described two types of models in detail: (i) Time-independent models with deterministic traveller knowledge and deterministic networks — called the Deterministic Network Deterministic User Equilibrium (DN-DUE) and (ii) time-independent models with stochastic traveller knowledge and deterministic networks — known as the Deterministic Network Stochastic User Equilibrium (DN-SUE). Within this thesis, the DN-DUE model was applied to two sample networks. The performance of different traffic assignment models in analyzing congestion relief measures was also discussed.
Results
The use of traffic assignment models in analyzing congestion relief measures is promising: With the exception of a few measures, these models perform very well. From this thesis, we have concluded that the DN-SUE is the most appropriate model for analyzing congestion relief measures, because it maintains a good balance between reality and task difficulty. The Frank-Wolfe algorithm can be applied to this model, which is simple to implement and supplies good results. Further research is necessary when applying this algorithm to larger networks, since computation time then becomes a problem