参数有界约束下平差模型的一种新算法  

A New Approach for the Parameter-Bounded Adjustment Model

作　　者：陈仲儿 左廷英[1] 宋迎春[1] 

机构地区：[1]中南大学地球科学与信息物理学院,长沙市麓山南路932号410083

出　　处：《大地测量与地球动力学》2015年第3期440-444,共5页Journal of Geodesy and Geodynamics

基　　金：国家自然科学基金(49774209)

摘　　要：基于不确定度理论,建立参数有界约束下的平差模型。为克服不等式约束模型单边约束的缺陷,从最小二乘出发,将问题转化为附有箱型约束的二次规划问题,提出一种求解参数最优估值的新算法,并给出参数估计的统计性质。数值实验表明,新算法简单、可行,具有较快的收敛速度,并能够在一定程度上减少部分数值的部分不确定性。Based on uncertainty theory, this paper establishes the parameter-bounded adjustment model and o- vercomes the defects of inequality-constrained adjustment. Due to the least squares, the problem is converted into a quadratic programming problem with box constraints, proposes a new algorithm to calculate the optimal values of parameters, and gives the statistical properties of parameter estimation. Numerical experiments vali- date that the new algorithm is simple, feasible, has a faster convergence speed, and show that the algorithm can reduce some numerical uncertainty.

关 键 词：不确定度理论 平差模型 参数有界 最小二乘估计 不等式约束平差 

分 类 号：P207[天文地球—测绘科学与技术]