Traffic congestion is an escalating phenomenon with multidimensional impact. The proposed research stands in the hypothesis that a traffic forecasting method, acknowledging several exogenous parameters for a more plausible representation of traffic evolution, with the implementation of corresponding ATM strategies, could restrain or prevent network’s performance deterioration.