基于遗传规划的炸弹弹道分组拟合模型

针对传统弹道拟合方法需要事先设定拟合方程的函数结构形式且不能很好适应大范围投射条件的问题,提出了一种分组遗传规划的炸弹弹道拟合方法。首先,依据动力学仿真实验得到炸弹在不同投射条件下的弹道样本。然后,采用分组遗传规划对样本数据中弹道诸元和投射初值进行输入输出映射拟合训练,得到了多组弹道拟合方程,并进行了未分组和分组拟合方程式之间的精度和泛化性比较。大量检测验算说明,采用分组遗传规划得到的弹道拟合方程式组具有物理关系明确、拟合精度高、泛化好、实时解算快等特性。

Trajectory fitting model of bomb based on group genetic programming

The traditional trajectory fitting method always needs a priori enactment about the structure form of the fitting equation and is not adaptable when it is used to fit trajectory under the large range launching condition.In view of these deficiencies, a bomb trajectory fitting method based on group genetic programming (GGP) is proposed.First, the GGP model samples data of different launching conditions are obtained through bomb dynamics simulation. Then, on the basis of mapping input output training of these trajectory data and initial launching value by the GGP, multiple trajectory fitting equations are provided. The accuracy between ungroup and group equations is compared based on modeling errors as well as generalization ability for untrained samples. A lot of checking calculation shows that the GGP method has clear physical relations, a high precision, good generalization and can fast calculate.