基于参数可变遗传算法的地面活动目标识别

特征维数压缩和神经网络隐含层结点数确定是雷达目标识别的关键问题，本文针对地面活动目标识别的特定问题构建了初始种群，然后提出一种即考虑进化代数对算法的影响，又考虑到每代不同个体适应度作用的可变交叉概率和变异概率，经过多次迭代进化最终收敛于问题的一个满意解。最后将该方法用于对三类地面活动目标的分类和识别，验证该方法用于目标分类和识别的可行性和有效性。

Ground Moving Targets Recognition Based on Alterable Parameter Genetic Algorithm

The target feature selection and the hidden layer numbers confirm of the neural network are the key problems in the radar target identification. In this paper initial population suitable for ground moving targets recognition is constructed. Then the alterable crossover probability and alterable mutation probability are proposed, which consider the influence of every generation to algorithm and the effect of different individual fitness in every generation. A satisfying result is acquired after many times crossover and mutation operation. At last by analysis and comparison of the final classification results of three kinds of ground moving targets, the feasibility and effectiveness are tested.