张世辉, 金同清, 张运杰, 周锐, 冉华明, 周礼亮. 群体智能专辑:基于自组织聚类的多机协同编批方法[J]. 工程科学学报. DOI: 10.13374/j.issn2095-9389.2023.10.09.002
引用本文: 张世辉, 金同清, 张运杰, 周锐, 冉华明, 周礼亮. 群体智能专辑:基于自组织聚类的多机协同编批方法[J]. 工程科学学报. DOI: 10.13374/j.issn2095-9389.2023.10.09.002
Multi-aircraft Collaborative Batching Method Based on Self-organizing Clustering[J]. Chinese Journal of Engineering. DOI: 10.13374/j.issn2095-9389.2023.10.09.002
Citation: Multi-aircraft Collaborative Batching Method Based on Self-organizing Clustering[J]. Chinese Journal of Engineering. DOI: 10.13374/j.issn2095-9389.2023.10.09.002

群体智能专辑:基于自组织聚类的多机协同编批方法

Multi-aircraft Collaborative Batching Method Based on Self-organizing Clustering

  • 摘要: 针对多机协同对抗过程中的编批问题,设计了一种基于改进自组织迭代聚类的多机协同编批方法。该方法解决了传统自组织迭代聚类算法中人工参数设置不便利不直观的问题,能够在给定少数直观超参数条件下,给出合理的编批结果。首先对高维多机态势信息进行标准化和主成分分析处理;然后引入密度聚类中的密度判据思想对传统自组织迭代聚类方法的合并和分裂操作进行改进,优化了分裂和合并操作所涉及的超参数;最后选取算法评价指标,对人工数据和实际场景数据进行聚类并对结果进行评价。仿真表明,改进自组织迭代聚类算法在人工数据以及真实场景中都具有较好的性能。

     

    Abstract: Aiming at the batching problem in the cooperative antagonistic processes of multi-aircraft, a multi-aircraft collaborative batching method based on improved self-organizing iterative clustering is designed. This method addresses the problem of inconvenient and non intuitive manually parameter settings in traditional self-organizing iterative clustering algorithms, and is capable of providing reasonable batching results under the condition that a few intuitive hyperparameters were given. . Firstly, the high-dimensional multi-aircraft situational information is standardized and subjected to principal component analysis. Then, the density criterion from density clustering is introduced to enhance the merging and splitting operations in the traditional self-organizing iterative clustering method, thereby optimizing the hyperparameters involved in these operations. Finally, algorithm evaluation metrics are selected to assessment the cluster results under artificial data and real-world scenario data. Simulation results demonstrate that the improved self-organizing iterative clustering algorithm performs well both on artificial data and in real-world scenarios.

     

/

返回文章
返回