摘要
文中采用二次规划的方法解决秩亏网平差时法方程的解不唯一的问题,提出了基于罚函数的二次规划秩亏网平差算法。研究结果表明传统的附加条件法是该算法的特例,通过最小二乘与最小范数原理筛选出的最优解精度更高。
In the paper, using quadratic programming method to solve the problem that the solution of the normal equation was not unique when carrying out rank-defect adjustment, a rank-defect adjustment algorithm was proposed based on quadratic programming with penalty function. Results showed that the additional restrain condition method was a special case of this algorithm. Precision of the optimal solution screened by using the least square and the minimum-norm principle was higher.
作者
柳佳彬
左廷英
Liu Jiabin;Zuo Tingying(Central South University,Changsha 410083,China)
出处
《矿山测量》
2018年第4期73-77,共5页
Mine Surveying
基金
国家自然科学基金(项目名称:基于复杂先验信息的大地测量不适定问题研究
编号:41674009)
关键词
秩亏网平差
二次规划
罚函数
惩罚因子
优化理论
rank-defect adjustment
quadratic programming
penalty function
penalty factor
optimization theory
