Forecast of Non-Equal Interval Track Irregularity Based on Improved Grey Model and PSO-SVM

Abstract

Effective forecasting of track irregularity development trend is critical to the railway maintenance and management. In China, track irregularity is commonly evaluated via an index called track quality index (TQI). According to the trend and random of the TQI development, a hybrid forecasting method composed of the improved grey model (GM(1,1)) and a particle swarm optimization support vector machine (PSO-SVM) is proposed in this paper. First, a non-equal interval sequence is transformed into an equal interval sequence, and an improved GM(1,1) based on the compound trapezoid quadrature formula is proposed to estimate the short-term rough trend of the TQI. Then, PSO-SVM is adopted to correct the rough results and obtain an accurate TQI trend. Numerical experiment results show that the proposed forecast method outperforms the track irregularity trend component and the grey nonlinear periodic correction method and the non-equal interval weighted grey model and neural networks in terms of relative error.

Keywords

• Track quality index
• non-equal interval forecast
• grey model
• support vector machine
• PSO