Expressway Pavement-Performance Indexes PredictionBased on Gray Model GM(1,1,λ)

摘要

The gray model GM(1,1,λ) method was applied to fit the pavementrnperformance indexes (international roughness index, pavement deflection value ,rnBPN)of asphalt expressway. To obtain the optimal value of parameter λ, the randomrnsearch method and the cyclic search method were established by minimizing thernmean absolute error, the average relative error or the average square of AbsoluternError. The results show that forecast precision of gray model GM(1,1,λ) method isrnhigher than that of GM(1,1) method. The optimal value of parameter λ can bernobtained easily and accurately by random search method and the cyclic searchrnmethod. The parameter λ is very close to 0.5 by minimizing the average square ofrnAbsolute Error. To improve the forecast precision，it is more appropriate to minimizernthe mean absolute error or the average relative error.