基于遗传算法的极短弧定轨(续)

Genetic Algorithm for Initial Orbit Determination with Too Short Arc(Continued)

采用遗传算法解决极短弧定轨问题时,由于遗传算法不同于经典方法的计算过程,野值剔除方法不再适用.在遗传算法中通过在适值函数中采用不同损失函数实现了稳健估计,解决了极短弧定轨中的野值处理问题.在遗传算法中不同损失函数的引入较经典方法大大简化.通过对多种损失函数的计算比较,表明采用最小中值二乘(LMS,Least Median Square)和截尾最小二乘(LTS,Least Trimmed Square)估计可大幅度提高极短弧定轨的稳健性,具有极高的崩溃点.

When using the genetic algorithm to solve the problem of too-short-arc（TSA） determination, due to the difference of computing processes between the genetic algorithm and classical method, the methods for outliers editing are no longer applicable. In the genetic algorithm, the robust estimation is acquired by means of using different loss functions in the fitness function, then the outlier problem of TSAs is solved. Compared with the classical method, the application of loss functions in the genetic algorithm is greatly simplified. Through the comparison of results of different loss functions, it is clear that the methods of least median square and least trimmed square can greatly improve the robustness of TSAs, and have a high breakdown point.