A Traffic Flow Forecasting Method for Microscopic Simulation

Abstract

The present invention relates to a traffic flow prediction method for microscopic simulation. The method is characterized by comprising the following steps that: a road network model is constructed, and microscopic simulation calibration is performed through using actual data; actual flow is collected, predicted flow data are obtained on the basis of the traffic flow prediction of a particle swarm optimized BP neural network algorithm, the predicted flow is outputted, so that simulation operation is performed, a simulation result in the simulation period is outputted, and the flow values of each road section are obtained; and with the simulation flow values adopted as input, prediction is performed through the prediction algorithm on the basis of the actually acquired flow until the difference values of the flow values obtained through the simulation and flow values obtained through the prediction algorithm satisfy pre-set requirements, and then predicted flow is evaluated in micro simulation software, so that the influence of the prediction result on future traffic can be researched, and analysis can be more scientific and reasonable.

Description

technical field [0001] The invention relates to a traffic flow prediction method for microscopic traffic simulation, belonging to the technical field of traffic prediction analysis. Background technique [0002] For traffic flow forecasting, many scholars at home and abroad have done a lot of research and obtained many forecasting methods, such as historical data average method, Kalman filter theory, artificial neural network method, etc. Different methods have their own advantages and disadvantages. The invention adopts the BP neural network optimized by particle swarm optimization. [0003] Due to the development of data acquisition methods in recent years, artificial neural network has become a research hotspot again. In terms of prediction, compared with other methods, the artificial neural network has strong robustness, adaptive nonlinear characteristics, and the ability of distributed information storage and parallel information processing, which is very suitable for ...

AI Technical Summary

Problems solved by technology

However, due to the rapid convergence effect of the particle population, it is prone to premature convergence.

Images