采用DDPG算法的弹道导弹突防诱饵分布空域

求解弹道导弹突防诱饵最佳分布空域，对导弹突防过程进行建模，建立导弹与诱饵运动模型、拦截器导引头识别模型以及突防效果评价模型。利用深度确定性策略梯度（deep deterministic policy gradient， DDPG）算法，仿真分析多种诱饵数量情况下的不同空域分布对突防效果的影响。结果表明，弹道导弹存在突防效果有效的空域分布，诱饵数量越多，突防效果越好。但是，随着诱饵数量的增加，突防效果的增强也会减弱，诱饵分布在导弹周围x≤10 km、y≤10 km、z≤10 km的空域范围内，导弹具有有效的突防效果，验证了模型的可行性。

Decoys Based on DDPG Algorithm

to find the optimal distribution of decoys for effective ballistic missile penetration,firstly,the missile penetration process is modeled into the motion model of missile and decoy,the identification model of interceptor seeker and the penetration effect evaluation model.Through the simulation of DDPG（deep deterministic policy gradient),the effects of different spatial distributions by a various number of decoys on penetration are analyzed.The results show that there exists an effective spatial distribution for the missile penetration.The more the number of decoys,the better the penetration effect.However,the enhancement of penetration efficiency will be weakened with the increase of the number of decoys.The missile has an effective penetration effect when the decoys are distributed in the airspace of x≤10 km,y≤10 km and z≤10 km around it,which verifies the feasibility of the model.