基于1stOpt软件的二元立木材积方程的研究  

Study on binary standing volume equations based on 1stOpt software

作　　者：杜婧[1] 冯仲科[1] 樊潇飞 张巍巍[1] 张凝[1] 

机构地区：[1]北京林业大学,北京100083 [2]西南大学,重庆400715

出　　处：《中南林业科技大学学报》2014年第4期64-67,共4页Journal of Central South University of Forestry & Technology

基　　金：国家科技支撑计划项目(2012BAH34B01);"亚太森林恢复与可持续管理网络"旺业甸林场多功能林业试点示范项目

摘　　要：选取旺业甸实验林场201株落叶松样木胸径、树高、材积精测数据,利用1stOpt优化分析计算软件平台的Levenberg-Marquarat+通用全局优化算法(LM-UGO)拟合二元材积方程,并与前人研究得到较好结果的遗传算法(GA)和加权最小二乘法(WLS)进行比较。结果表明,LM-UGO、GA和WLS均可得到较高精度的二元材积方程,LM-UGO方法能够比GA和WLS法更好地拟合二元材积方程,且具有操作方便、拟合快速而准确、收敛速度快等优点,是拟合二元立木材积方程的有力工具。经F检验,三种方法的F值LM-UGO<GA<WLS<F0.05(2,19),且TRB和MSB值均小于±3%,满足林业调查技术的要求。The binary standing volume equations were studied by the 201 samplings of larch based on the field precision measuring data of the DBH, the height and the volume in Wangyedian Experiment Tree Farm. The volume equations were established by using Levenberg-Marquardt（LM-UGO）method of 1stOpt software, and were compared with Genetic Algorithm method and Weighted Least Square method that both had achieved better results in previous researches. The results show that the binary standing volume equations were established by using LM-UGO, GA and WLS, but the LM-UGO method can better fit the binary standing volume equation than GA and WLS;the LM-UGO method of 1stOpt software with easier operation, more rapid and accurate fitting, higher convergence speed, etc., is a powerful tool for fitting binary standing tree volume equation. According to F-test, the F values of the three method ranked as the following order：LM-UGO〈GA〈WLS〈F0.05（2,19）, while the TRB and MSB values of all them were less than ±3%, meeting the requirements of forest survey technology.

关 键 词：森林 落叶松 二元材积方程 1stOpt软件 通用全局优化算法 遗传算法 加权最小二乘法 

分 类 号：S711[农业科学—林学]