基于自编码神经网络的装备体系评估指标约简方法

来源期刊:中南大学学报(自然科学版)2013年第10期

论文作者:张乐 刘忠 张建强 任雄伟

文章页码:4130 - 4138

关键词:武器装备体系;作战效能评估;Autoencoder网络;深度学习

Key words:weapon system-of-systems; combat effectiveness assessment; Autoencoder; deep learning

摘    要:基于现代战争的信息化程度、复杂性和不确定性不断提高,对武器装备体系(WSoS)作战效能的评估提出了更高的要求。为准确评估系统的作战效能,建立合理的作战效能评估指标体系非常关键,然而,影响作战效能的指标繁多,指标间存在冗余和相关性,直接采用这些指标会增加后续效能评估的时空复杂度,在创建作战效能评估指标体系基础上,采用Autoencoder网络深度学习(Deep learning)方法实现评估指标集合的约简化,将复杂的指标体系非线性映射到低维的指标数据中,从而明显减少数据的维数,保留关键重要的指标,去除冗余的指标。实验结果表明:约简后的指标能够很好地代表原有的指标数据,从而明显降低后续作战效能评估工作的计算复杂度。

Abstract: As the information degree, complexity and uncertainty of modern war increase, effectiveness assessment for weapon system-of-systems (WSoS) from the sea is more important than ever. Constructing appropriate evaluation index for WSoS is critical to evaluate its effectiveness, affecting the operational effectiveness of a range of indexes, indexes redundant and direct use of these indexes will increase the time-space complexity of effectiveness assessment, multiple and multi-scale indexes for the system were established. In order to reduce the complexity of index, Autoencoder was taken to project these data into low dimensional data non-linearly. The results show that the reduced data can exactly represent the origin data when the original dimension is reconstructed. This will highly reduce the complexity of the next assessment process of WSoS.

相关论文

  • 暂无!

相关知识点

  • 暂无!

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号