Research on Short-Term Traffic Flow Forecast Based on Improved Cuckoo Search Algorithm

Abstract

In recent years, the traffic flow in big cities has increased significantly, and the development of intelligent transportation systems (ITS) has become the general trend. Most of the traditional traffic flow prediction models are highly dependent on experienced experts and lack the ability to learn independently. Since the traffic flow depends on various factors such as weather, road conditions, and whether there are major events, it’s influenced by multi-factors and huge amounts of data. It is difficult to fit and process traffic flow data well, so traditional traffic forecasting models are no longer suitable for the analysis and prediction of current and even next-generation traffic flow. Now artificial neural networks are widely used in traffic flow forecasting due to their strong robustness and fault tolerance, high operating efficiency, ability to process massive data, strong nonlinear mapping capabilities, and strong learning and adaptive capabilities. To tackle the problem of random initial weights and thresholds of the traditional model for the neural network, we proposes an improved cuckoo search algorithm, and uses the improved algorithm to optimize the initial threshold in the long short-term memory neural network, and applies the improved model to the research on traffic flow prediction, in order to improve traffic planning and save people’s travel time and fuel costs laid the groundwork.