A New Approach for the Prediction of Traffic-Induced Ground Vibration Using a Hybrid Optimized Anfis-Based Model

Abstract

An attempt has been made to evaluate and predict the traffic-induced ground vibration using a hybrid optimized ANFIS-based model. Towards this aim, ground vibrations caused by traffic were monitored on a building located near the road. To investigate the appropriateness of this approach, the prediction by ANFIS was also compared with the most widely used vibration predictors. In this research, a hybrid of adaptive neuro-fuzzy inference system (ANFIS) optimized by particle swarm optimization (PSO) and genetic algorithm (GA) was proposed to predict traffic-produced ground vibration. The performance criterion selected for the comparison between the actual and the predicted data were the sum of squares due to error (SSE), the root mean square error (RMSE), and goodness of fit (R-square, adjusted R-square). It turns out that the hybrid GA-ANFIS prediction model outperforms the commonly used predictors and conventional ANFIS.