论文部分内容阅读
文中针对海洋环境影响下单武器装备作战效能的评估问题,建立基于RBF神经网络的评估模型。在实际应用中,为了保证评估结果的客观性和准确性,提出一种基于统计原理的改进RBF神经网络模型。该改进模型采用基于样本相似度的聚类算法,以加权欧氏距离为样本相似性度量方法,通过对样本进行聚类处理得到RBF神经网络模型的参数,进而建立评估模型。最后,为了验证提出模型的可行性,利用样本实例对模型进行训练,并利用训练后的模型对某一环境下单一武器作战效能进行评估,实验结果表明了模型的可行性和可靠性。和传统方法相比,该评估模型基于样本数据的统计信息,不需要专家知识,具有较高的客观性。
In this paper, aiming at the evaluation of combat effectiveness of single weapon equipment under the influence of the marine environment, an evaluation model based on RBF neural network is established. In practical application, in order to ensure the objectivity and accuracy of the evaluation results, an improved RBF neural network model based on statistical principles is proposed. The improved model uses the clustering algorithm based on sample similarity and the weighted Euclidean distance as the sample similarity measure method. The parameters of the RBF neural network model are obtained by clustering the samples, and the evaluation model is established. Finally, in order to verify the feasibility of the proposed model, the sample is used to train the model, and the trained model is used to evaluate the combat effectiveness of a single weapon in an environment. The experimental results show the feasibility and reliability of the model. Compared with the traditional method, the evaluation model is based on the statistical information of the sample data, does not require expert knowledge, has a high objectivity.