【摘要】 分别采用遗传算法参数估计方法与遗传神经网络方法 ,并通过对实测数据进行的分析 ,建立了路面性能综合评价方法。在参数估计方法中 ,首先建立专家综合评价与道路实测数据之间的回归模型 ,通过对模型进行遗传算法参数优化 ,得到了相关度很高的结果。同时 ,还应用遗传神经建立主观综合评价和客观实测数据之间的高度非线性关系 ,避免了 BP网络的缺陷。结果表明 ,该模型具有很好的相关性。两种方法为路面管理系统路况分析、路面养护评价以及路面养护对策的排序提供了重要依据更多还原

【Abstract】 Analyzed the practical data by two genetic algorithms based method,genetic algorithm parameter estimation method and genetic algorithm neural network nonlinear method,this article raised two pavement performance synthetic evaluation models.In the parameter estimation method,firstly,a mathematic regress model is established.A high correlation coefficient of this model is solved by genetic algorithm parameter estimation method.Meantime,genetic algorithm neural network nonlinear method that avoids the BP network’s defects is severed in establishing a high nonliear relationship between subjective synthetic evaluations and objective practical data.The result shows that this model have higher correlation coefficient.These methods provide a basic reference for pavement performance analysis,maintenance decision sorting and evaluation,and make an availability discuss on the application of genetic algorithm in pavement performance synthesis evaluation.更多还原