基于灰狼算法与小波神经网络的目标威胁评估

为了提高目标威胁度评估的精确度，建立反向学习灰狼算法（OGWO）优化小波神经网络的目标威胁评估模型（OGWO-WNN），提出基于该模型的算法.该模型使用反向学习策略（OBL）优化灰狼算法（GWO），通过改进后的灰狼算法优化小波神经网络的各权值和小波基函数的平移因子与伸缩因子，使优化后的小波神经网络能够对威胁度测试样本集作更好的预测.实验结果显示，采用反向学习灰狼算法能够更好地优化小波神经网络的权值与平移、伸缩因子，使建立的小波神经网络目标威胁评估模型具有更高的预测精度和更强的泛化能力，能够精准、有效地实现目标威胁评估.

Target threat assessment using grey wolf optimization and wavelet neural network

Opposition-based learning grey wolf optimization (OGWO) and wavelet neural network model (OGWO-WNN) was established in order to improve the air targets threat estimation accuracy, and the algorithm based on the model was proposed. Opposition-based learning (OBL) was adopted to optimize grey wolf optimization (GWO) algorithm, and OGWO was employed to simultaneously optimize the weights and translation and scalability factors in WNN. Target threat database was adopted to test the performance of OGWO-WNN in target threat prediction. The experimental results show that the weights and translation and scalability factors in WNN can be better optimized. The air target threat assessment model has higher prediction precision and better generalization ability, and can accurately and effectively complete target threat estimation.