摘要
Lgistic是一种近"S"型的曲线生长方程,其3个参数不能用常规的最小二乘法来求解。该文应用遗传算法的实数编码方案,利用Matlab 6.5和英国设菲尔德大学开发的遗传算法工具箱函数编制了参数拟合程序。引用了周元满等(2004)的4种不同桉树品系的平均蓄积量原始数据,进行生长过程拟合,其结果与原作者所求解的参数相比较,残差平方和均相对偏小,其中尾叶桉的相差最大为289.38,EC_1品系相差最小为74.07,其模型求解精度都有所提高。
Logrsue equation fits near S-shapea growth curves,but the solution of its tnree parameters cannot be obtained by traditional least square algorithm.The real-coding scheme in genetic algorithm was applied and parameter fitting programme was established in this paper using Matlab 6.5 as well as genetic algorithm tool functions developed by the University of Sheffield,UK.The growth process of four eucalyptus varieties was simulated according to the primary data of average stock volume derived from the publication of Zhou et al (2004).The results show that,compared with the findings of Zhou et al (2004),the residual sum of square is lower,with the largest difference in Eucalyptus urophylla,289.38,and the least in EC_1 variety,74.07. The precision of this model is improved to some extent.
出处
《北京林业大学学报》
CAS
CSCD
北大核心
2008年第S1期192-195,共4页
Journal of Beijing Forestry University
基金
北京市教育委员会共建项目建设计划
湖北民族学院博士启动基金项目
