中文摘要：

空间目标的巡天观测获取了海量的极短弧观测数据，而经典初轨计算方法对于极短弧几乎不能获得合理的结果．将初轨计算问题转换为两个三变量的分层优化问题，采用遗传算法，针对具体问题选择了优化变量以及相应的遗传操作，建立了一种极短弧初轨计算方法．基于实测资料的数值实验表明，方法可为后续工作提供有效的初值．

英文摘要：

The sky surveys of space objects have obtained a huge quantity of too- short-arc （TSA） observation data. However, the classical method of initial orbit deter- mination （IOD） can hardly get reasonable results for the TSAs. The IOD is reduced to a two-stage hierarchical optimization problem containing three variables for each stage. Using the genetic algorithm, a new method of the IOD for TSAs is established, through the selection of optimizing variables as well as the corresponding genetic operator for specific problems. Numerical experiments based on the real measurements show that the method can provide valid initial values for the follow-up work.