Short-Term Traffic Flow Forecasting of Intersection Based on Approximate Dynamic Programming

Abstract

In this paper, we propose a model for short-term traffic flow forecasting based on approximate dynamic programming when considering the influence of related roads. First, construct the system performance index using the error function as the objective function, adjust the action network online according to the minimum index to generate the approximate optimal control vector and the control matrix, then we can get the traffic flow data at the specific road and time to forecast. To illustrate how this method works, we use intersection traffic flow of Hongli road and Shangbu road in Futian Shenzhen as experimental data, which can be demonstrated that this method can increase the forecasting accuracy and meet the real-time forecasting requirements.