摘要
冻土融沉系数是冻土融化后沉降量的直接反映,影响融沉系数的主要因素有粉黏粒含量、未冻土的含水率、干密度、液塑限等.为了定量化的描述融沉系数与各影响因素之间的关系,引入基因表达式编程算法进行五维空间上的函数挖掘,将函数表达式基因化,通过Karva语言对基因进行编码,利用达尔文的进化理论,初始种群经过若干代的交叉变异,最终得出了最优化的个体,即最优函数表达式.结果表明:得到的函数表达式能综合反映各因素与融沉系数之间的关系;对测试样本进行预测,R2值为99.50%,实现了融沉系数预测的高精度和直观化,为融沉系数预测提供了一种新的智能化方法.
The thaw-settlement coefficient is a direct reflection of the settlement after thawing, the main factors that affect the coefficient are powder clay content, moisture content of not frozen soil, dry density, liquid and plastic limit. To quantitative descript the financial relationship between the thaw coefficient and its influencing factors, introduce gene expression programming algorithm to find a optimal function in the five-dimensional space, gene which means function code by Karva language in the computer, learn from Darwin＇s theory of evolution, after several generations, crossover and mutation, the initial population ultimately evolutes＇ an optimal individual, namely the optimal function. The results showed that the function can be integrated to reflect the relationship between the factors and the coefficient of thaw. To predict the test samples, R2 is 99. 500/00, achieve coefficient prediction precision and intuitive. It＇s a new intelligent method for the thaw-settlement coefficient prediction.
出处
《宁夏大学学报（自然科学版）》
CAS
2013年第2期147-150,共4页
Journal of Ningxia University(Natural Science Edition)
基金
国家自然科学基金资助项目(50808077)
"中央高校基本科研业务费"资助项目
关键词
基因表达
冻土
融沉系数
预测
复相关系数
gene expression
permafrost
thaw-settlement coefficient
forecasts
multiple correlation coefficients
