Hybrid Genetic Algorithm and time delay neural network model for Forecasting Traffic flow

摘要

The world is gradually moving towards being a smart planet. Intelligent Transportation Systems are a key constituent of any smart planet solution. Forecasting Traffic flow is one of the aspects of Intelligent Transportation Systems. A hybrid model comprising of genetic algorithm and time delay neural network has been used here to predict short term traffic flow. Three new input parameters; dew point, humidity and sea level pressure have been incorporated into a multi input parameters model. The hybrid model leads to significant reduction in RMSE values as compared to neural network model. Hence, combination of multi input parameter models with Genetic Algorithm has the potential to improve the accuracy of prediction of short term traffic flow.