基于共轭梯度法和最速下降法的非线性测量数据处理  

Nonlinear Survey Data Processing Based on the Conjugate Gradient Method and the Steepest Descent Method

作　　者：宁伟[1] 卿熙宏[2] 陶华学[2] 

机构地区：[1]山东农业大学信息科学与工程学院,山东泰安271018 [2]山东科技大学地球信息科学与工程学院,山东泰安271019

出　　处：《山东科技大学学报（自然科学版）》2004年第4期5-7,共3页Journal of Shandong University of Science and Technology(Natural Science)

基　　金：国家自然科学基金资助项目 (40 1 740 0 3 )

摘　　要：将共轭梯度法与最速下降法有机结合起来,构造出一种解决非线性测量数据处理问题的新方法———混合算法。这种方法充分利用了共轭梯度法和最速下降法良好的收敛优点,既提高了共轭梯度算法的收敛速度,又解决了目标函数"性态不优"时,最速下降法难以解决的问题。文中的算例结果表明,混合算法与单纯的共轭梯度法或最速下降法相比,具有收敛速度快、收敛范围大、适应面宽等特点。The conjugate gradient method and the steepest descent method are combined, and a new method-mixed algorithm method for solving the problem of nonlinear survey data processing is created in this paper. The mixed method makes use of their good convergence merit, and raises the convergence rate of the conjugate gradient method and solves the problem for which the steepest descent method can not solve in the condition with bad characteristics for objective function. Compared with the conjugate method or the steepest descent method, the method has features with quick convergence rate, large convergence range and wide accommodation.

关 键 词：非线性数据处理 共轭梯度法 最速下降法 混合算法 

分 类 号：P207[天文地球—测绘科学与技术]