海洋环境下武器装备作战效能自学习评估模型

摘要

A self-learning evaluation model is proposed to evaluate the operational effectiveness of weapon equipment under the influence of marine environment. By dividing operational effectiveness into different levels,nonlinear mapping between marine environmental ele-ments and the levels of operational effectiveness is established based on neural network. Then the operational effectiveness can be obtained by evaluating the operational effectiveness level of the weapon equipment. To improve the reliability of the evaluation,the model is re-vised by adding the ability of self-learning. With the increase of samples,operational effectiveness can be further divided into more levels to improve the numerical precision of the evaluation result,which results in improvement of the structure of neural network in evaluation model. On the basis of that,the model can be updated through self-training of the model. Finally,aiming at the operational effectiveness evaluation instance of an operation platform,the experimental results show that the reliability of the evaluation model is raised through self-learning,which proves the feasibility of the proposed model. Compared with the traditional methods,the model proposed needs no ex-pertise,which makes the evaluation results more objective.%　　针对海洋环境影响下武器装备作战效能的评估问题，提出了具有自学习能力的评估模型。将作战效能分级，基于神经网络建立海洋环境要素与作战效能级别之间的非线性映射，通过估计武器装备的作战效能级别获得其作战效能。为提高模型的评估可信度，评估模型增加了自学习能力，可对其本身进行自学习修正。随着样本增加，为提高评估结果的数值精度，对作战效能级别作进一步细分，使得模型中神经网络的结构得以改进。在此基础上，通过对模型的自学习训练即可实现模型的自学习修正。最后针对某作战平台作战效能的评估实例，实验结果表明模型自学习更新后其可信度得到提高，从而验证了所提出的评估模型的可行性。与传统方法相比该评估模型无需依赖专家经验，具有较高的客观性。