The paper studies the effects of control and prediction horizons on a multimodal traffic regulation strategy. The strategy acts on traffic lights in order to regulate Private Vehicle (PV) traffic by minimizing their number in the network and to regulate the Public Transport (PT) traffic by maximizing their schedule adherence. The strategy architecture is based on an automatic closed loop and uses the principles of the General Model predictive Control (GMPC). The GMPC defines two blocs : the prediction bloc and the optimizer. The predictor is based on two models belonging to PV and PT traffic. These two models balance simplicity and accuracy. The optimizer is based on the Particle Swarm Optimization, a powerful metaheuristic, that resolves a bi-objective optimization problem. The problem, as it is formulated within GMPC architecture, evolves control and prediction horizons. The simulation study focuses on these two important parameters and their effects on the traffic regulation efficiency.