基于GEP算法的沥青混合料动模量预测

Prediction of Dynamic Modulus of Asphalt Mixture Based on Gene Expression Programming Algorithm

关键词:  道路工程  沥青混合料  基因表达式编程算法  动模量  预测模型

Key words:  road works  asphalt mixture  gene expression programming(GEP) algorithm  dynamic modulus  prediction model

摘要:

采用基因表达式编程算法对沥青混合料动模量进行预测.以沥青混合料空隙率（Va）、有效沥青含量（wbeff）、沥青黏度（η）、荷载频率（f）、集料筛余质量分数（ρ34，ρ38，ρ4）以及集料在0075mm筛孔上的通过率（ρ200）为主要参数，建立了基于基因表达式编程算法的沥青混合料动模量预测模型.结果表明：由预测模型得到的动模量预测值与实测值之间具有较高的相关性；将预测模型与Witczak 1999模型、韩国动模量预测模型和人工神经网络模型等方法进行比较后发现，采用基因表达式编程算法来预测沥青混合料动模量具有简单可靠的优点.

Abstract:

Gene expression programming(GEP) algorithm was used to predict dynamic modulus of asphalt mixture, with eight main factors of dynamic modulus, i.e. air void（Va），effective binder content（wbeff）, viscosity of binder（η），loading frequency（f），the mass fraction of aggregate sieve residue on the sieve size of 19（ρ34），95（ρ38），475mm（ρ4） and aggregate passing rate by the 0075 mm sieve（ρ200）, which constitute the main factors to predict dynamic modulus model of asphalt mixture. GEP algorithm can be applied to establish a dynamic modulus prediction model of asphalt mixture by discrete 8 factors. The results show that between the dynamic modulus predicted and measured values a high correlation is obtained and by compared with Witczak 1999 function model, Korean dynamic modulus prediction model and artificial neural network model, GEP algorithm prediction model has some of superiority over the other models.