The User Equilibrium Traffic Assignment Method Under Interval Uncertain Demand

It is difficult to make accurate predictions of future travel demand due to many sources of error and uncertainty which are paid few attention in traditional forecasts. Meanwhile, most present travel forecasting models produce only a single answer, which is usually far from the actual value, leading to higher decision-making risks. Among the four-step forecasting models, the user equilibrium assignment is the most challenging job. Applying the theory of uncertainty, the author presents a method to solve the UE assignment problem under uncertain demand.The problems about user equilibrium traffic assignment models and algorithms are studied in this thesis. Firstly, the uncertainty of traffic demand is analyzed to point out the main factors clearly. Current uncertainty measure methods are compared to indicate that the interval measurement of uncertain traffic demand is more superior. Then, aiming at the interval uncertain demand, the route choice behavior of road users is analyzed. The influences to route choice behaviors which is from the subjective value of road users are considered, afterward the reliability based on interval travel time and impedance calculation methods are proposed, the user equilibrium assignment model under the conditions of interval uncertain demand is constructed, and the improved algorithm based on MSA is given. Finally, Using MATLAB, two simple road networks are utilized to analyze and verify the method.The problem of traffic assignment under uncertain demand is studied by application interval analysis method in this thesis, both the shortages of ignoring uncertain demand in traditional assignment methods and the influences of route choice behavior caused by the subjective of road user are all considered. There is uncertain information in the results of interval demand traffic assignment, and that rational consider limits are offered to decision-makers, which can improve the reliability of evaluation and reduce the risk of decision-making effectively. Therefore, the theoretical method is proposed in this thesis is significant to enrich the current traffic assignment theory, and is valuable for decision risk analysis of transport projects.