A Chance Constrained Model Of Traffic Flow Assignment

The aim of this thesis is to find the traffic flow assignment under the following factors:(i) travel time, increases with the demand; (ii) O-D demand, decreases with the travel time. On the basis of the classical flow assignment model, it is very hard for the demand to satisfy a new constraint exactly. Therefore, we establish the problem as a minimization problem where the O-D demand is a function of the travel time and satisfies a chance constraint. In order to find a reasonable flow assignment, we convert this traffic flow assignment problem to an equivalent minimization optimization problem. We show that the minimization problem is still a user-equilibrium (UE) problem, which can be reformulated as a second-order cone program (SOCP). And, SOCPs can be solved very efficiently by using interior point algorithms.This paper is organized as follows:In the third chapter, based on the classical user-equilibrium (UE) traffic flow assignment model, we discuss the user-equilibrium with variable demand model and the stochastic traffic flow assignment model, the symmetric or asymmetric link interactions in two-way traffic problem is also discussed. And then, we analyze the relations and differences between these four traffic flow assignment problem and UE problem.In the fourth chapter, we discuss and improve the flow assignment problems. The given example about four intersections is used to analyze the delay at intersection; and a new traffic flow assignment model is proposed, named as a chance constrained model of traffic flow assign-ment. The final example demonstrates that the new model is practical for solving traffic flow assignment problems.