CAJ | 학술논문

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。
문중침대해양배경영향하단무기장비작전효능적평고문제，건립기우RBF신경망락적평고모형。재실제응용중，위료보증평고결과적객관성화준학성，제출일충기우통계원리적개진RBF신경망락모형。해개진모형채용기우양본상사도적취류산법，이가권구씨거리위양본상사성도량방법，통과대양본진행취류처리득도RBF신경망락모형적삼수，진이건립평고모형。최후，위료험증제출모형적가행성，이용양본실례대모형진행훈련，병이용훈련후적모형대모일배경하단일무기작전효능진행평고，실험결과표명료모형적가행성화가고성。화전통방법상비，해평고모형기우양본수거적통계신식，불수요전가지식，구유교고적객관성。
An evaluation model based on RBF neural network is established to solve the evaluation problem of single weapon equipment operational effectiveness under the influence of marine environment. In practical application,to ensure that the evaluation result is objec-tive and exact,an improved RBF evaluation model based on principle of statistics is proposed here. The improved model uses clustering algorithm based on the sample similarity and utilizes weighted Euclidean distance as measure method of sample similarity,to get the pa-rameters of RBF neural network model by clustering process on sample data and then establish the model. Finally,to verify the feasibility of the proposed model,a set of actual sample data is used to train the model and use the trained model to evaluate the operational effec-tiveness of single weapon equipment under the influence of marine environment. The test has showed the feasibility and reliability of pro-posed model. Compared with the traditional methods,the evaluation model proposed is based on the statistics of sample data and needs no expertise,which makes the evaluation results more objective.