Highway traffic flow prediction method

Abstract

The invention discloses a highway traffic flow prediction method, and relates to the field of intelligent transportation. The highway traffic flow prediction method comprises the following steps: establishing a highway traffic flow prediction model based on an improved firefly-radial base function neural network; designing an improved firefly algorithm for training parameters of a radial base function neural network, thus increasing the accuracy and the convergence speed of the radial base function neural network; computing loss function value of the radial base function neural network, and selecting the radial base function neural network with the minimum loss function for predicting highway traffic flow. The highway traffic flow prediction method provided by the invention is fast in computation speed, is accurate in prediction accuracy and has a very good effect in highway traffic flow prediction.

Description

technical field [0001] The invention relates to the field of intelligent transportation, and more specifically, to a method for predicting traffic flow on expressways. Background technique [0002] With the continuous growth of social economy, more and more domestic cars are owned, and the traffic flow of expressways has increased sharply, which has led to more and more serious vehicle congestion on expressways. The existing method adopts the algorithm of radial basis function neural network to train the network parameters, which is easy to fall into the local minimum in the rough search process. Therefore, how to improve the stability of radial basis function (RBF) neural network for expressway traffic flow prediction is the key to the problem. [0003] The patent application with the application number 201710039355.5 "A Traffic Flow Prediction Method Based on Firefly Algorithm and RBF Neural Network" initializes the parameters of the firefly algorithm, initializes the fir...

AI Technical Summary

Problems solved by technology

[0003] The patent application with the application number 201710039355.5 "A Traffic Flow Prediction Method Based on Firefly Algorithm and RBF Neural Network" initializes the parameters of the firefly algorithm, initializes the firefly population with a random method, and encodes each individual in the population; uses the firefly algorithm to train RBF Neural network to obtain the optimal individual of the population; decode the optimal individual of the population to obtain a trained RBF neural network; use the trained RBF neural network to predict traffic flow data samples, this method has good predictive ability and Generalization ability, but there are still some deficiencies in training efficiency and prediction accuracy

Images