Title: The PSO optimisation SVM prediction model for the asphalt pavement environment and service fatigue life

Authors: Yu Sun; Dongpo He; Jun Li

Addresses: Road and Railway Engineering, Civil Engineering College, Northeast Forestry University, No. 26 Wo Hing Road, Xiangfang District, Harbin City, Harbin, Heilongjiang Province 150040, China; Department of Municipal and the Environmental Engineering, Heilongjiang Institute of Construction Technology, College Road, Limin Development Zone, Harbin 150000, China ' Road and Railway Engineering, Civil Engineering College, Northeast Forestry University, No. 26 Wo Hing Road, Xiangfang District, Harbin City, Harbin, Heilongjiang Province 150040, China ' Department of Municipal and the Environmental Engineering, Heilongjiang Institute of Construction Technology, College Road, Limin Development Zone, Harbin 150000, China

Abstract: In order to improve the accuracy of prediction by support vector machine (SVM), parameter optimisation of SVM is an important part of asphalt pavement life prediction. In this paper, a particle swarm optimisation support vector machine (PSO_SVM) method was proposed to predict the fatigue life of SBS modified asphalt mixture. This method combines SVM with particle swarm optimisation (PSO), makes full use of SVM's unique advantages in dealing with small sample regression problems and PSO global search optimisation, improves convergence speed, and achieves depth and breadth optimisation. Experimental results show that this method improves the parameter selection efficiency of SVM, and the prediction results are more accurate than those of ANN and SVM.

Keywords: particle swarm optimisation; PSO; support vector machine; SVM; SBS; modified asphalt mixture; environmental impact; fatigue life.

DOI: 10.1504/IJICT.2022.123173

International Journal of Information and Communication Technology, 2022 Vol.20 No.4, pp.355 - 366

Received: 10 Aug 2020
Accepted: 22 Sep 2020

Published online: 01 Jun 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article