参数带区间约束的子空间截断牛顿平差算法

基于积极集思想,利用子空间截断牛顿法提出参数带区间约束的平差问题算法。由于新算法可以迅速改变积极约束集的构成,其效率比不等式约束平差算法更高,并且可以对参数估计的精度进行评定。通过测边网算例说明,新算法能有效地降低模型的不适定性,保持参数的统计、几何或物理意义。

Subspace Truncation Newton Method for Parameter-Bounded Adjustment Problem

In the process of measuring data processing, the application of reliable prior information can effectively reduce the discomfort of the adjustment model. Based on active set thought, in this paper, using the truncated Newton method, we propose a new algorithm to solve the parameter-bounded adjustment problem. Our algorithm is more efficient than the inequality constraint, with its capacity to change the composition of the active set rapidly, and the precise evaluation of parameter value is given at the same time. Examples are given to show that the algorithm can effectively reduce the unfitness of the model and can maintain the statistical, geometric or physical significance of the parameters.