Abstract : This thesis deals with the development of dynamic traffic assignment models, an evaluation of their operational applications, and the supply-demand equilibrium of traffic. The analysis principally concerns Leurent's LADTA model (2003), which uses physical and economic assumptions close to those of static assignment while adding a chronological dimension and addressing transport link congestion using queues. Starting with an abstract expression of the analytical formulation of this model (which is generally applicable to dynamic assignment), we elaborate various algorithms using different mathematical formulations and endogenous variables. We discuss equilibrium algorithms and propose a hybrid algorithm which simultaneously takes into account link volumes and times. We give a formal analysis of the algorithm' s convergence, and we provide rigorous and computationally efficient convergence criteria. A simplified version of the model, its equilibrium algorithms, and their convergence criteria were then programmed. This prototype was applied to test cases in order to establish the behavior of the algorithms and criteria, and to permit the adjustment of certain parameters.