An Approach to Traffic Volume Forecasting Based on Ant Colony Neural Network

Combining the ant colony algorithm (ACA) and the neural network (NN), the present paper puts forward an approach to traffic volume forecasting based on the ant colony neural network. The approach employs the ACA with mutation features to train the weights of an artificial neural network (ANN), thus it is characterized by large mapping capacity of the NN, and by rapidity, global convergence, and heuristic learning of the ACA. Simulation results of an actual example indicate that this approach can improve the forecasting accuracy, hence its validity.