AHSBP algorithm and application in traffic parameters prediction

In order to avoid the low stability of memory and the uncertainty of global minimum point, a new algorithm based on harmony search that optimizes the weights and the thresholds of BP neural net is proposed to predict the traffic parameters. Then the algorithm is improved to increase the running spee...

Full description

Saved in:
Bibliographic Details
Published inChinese Control and Decision Conference pp. 4073 - 4077
Main Authors Qiong Wu, Xiangmo Zhao
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.05.2017
Subjects
Accidents
BP neural net
Convergence
Fitting
Harmony search
Linear programming
Neural networks
Optimization
Prediction algorithms
Signal processing algorithms
Traffic parameters prediction
Online AccessGet full text
ISSN1948-9447
DOI10.1109/CCDC.2017.7979213

Cover

More Information
Summary:In order to avoid the low stability of memory and the uncertainty of global minimum point, a new algorithm based on harmony search that optimizes the weights and the thresholds of BP neural net is proposed to predict the traffic parameters. Then the algorithm is improved to increase the running speed. The other advantage is that the algorithm combines the nonlinear fitting capability of BP neural net with the searching ability of global optimal solution of HS. And in this paper, the single parameter and the multiple parameters of traffic information are simulated to predict their tendencies. The results show that the proposed method has strong robustness, adaptability and global searching ability.
ISSN:1948-9447
DOI:10.1109/CCDC.2017.7979213