基于BP网络和序优化方法的火控攻击方案寻优

火控攻击方案寻优是一个基于仿真的优化问题，针对传统优化方法计算量大而且耗时多的不足，运用序优化理论通过排序比较和目标软化等思想来缩小搜索空间、提高搜索效率。首先建立了以导弹目标截获概率为效能指标的方案评估精确模型，然后用拟合后的基于误差反向传播（back propagation, BP）算法的神经网络作为粗糙评估模型，最后运用序优化搜寻达到一定目标截获概率的较优火控攻击方案。算例分析表明，序优化方法能在确保以足够高的概率寻得足够好方案的同时，有效提高搜索效率。

Fire control attack scheme optimization based on BP network and ordinal optimization

Fire control attack scheme optimization is one of the simulation optimization problems. Considering the large calculative load and a great time cost required for traditional optimization algorithms, the ordinal optimization theory is used to narrow down the searching space and improve the searching efficiency. Firstly, an accurate evaluation model which is on the basis of missile acquisition probability is established. Secondly, the neural network based on back propagation (BP) fitted is treated as the rough evaluation model. Finally, the ordinal optimization is used to search a better attack scheme which reaches a certain acquisition probability. The example analysis shows that the ordinal optimization can find a good scheme in high probability and improve the searching efficiency effectively.