基于因果熵的无人集群对抗评估指标分配方法
Index allocation method for unmanned swarm confrontation evaluation based on causal entropy
范波 1钟季龙 1徐丽霞 1吕筱璇 2王鹥喆 2刘禹 2侯新文2
作者信息
- 1. 军事科学院国防科技创新研究院,北京 100071
- 2. 中国科学院自动化研究所,北京 100190
- 折叠
摘要
针对传统无人集群对抗评估指标分配方法仅能选出相关性较好的评估指标,存在指标混淆和虚假相关问题,提出因果熵概念,用于度量评估指标和对抗任务完成效果间的确定性程度,从而滤除评估指标之间的混淆效应.以仿真环境下的无人集群空地协同围捕试验为例,通过评估指标因果分配和约简优选出更能反映无人集群对抗能力且相对独立的评估指标.仿真结果表明,所提方法可以有效地优化评估指标体系,获取更具代表性的评估指标,具有一定的普适性.
Abstract
In response to the problem of index confusion and false correlation in traditional unmanned swarm confrontation evaluation index allocation methods,which can only select indexes with good correlation,causal entropy is innovatively is proposed to measure the certainty degree between the evaluation indexes and the effect of the confrontation task,in order to filter out the confusion among evaluation indexes.Taking the unmanned swarm air-ground collaborative encirclement test in the simulation environment as an example,the evaluation indexes that better reflect the combat capability of the unmanned swarm which are relatively independent are selected and optimized through causal allocation and reduction.The simulation results show that the proposed method can effectively optimize the evaluation index system,obtain more representative evaluation indicators,and has a certain degree of universality.
关键词
结构因果模型/因果熵/指标分配/指标约简/冗余度Key words
structural causal model/causal entropy/index allocation/index reduction/redundancy degree引用本文复制引用
出版年
2024
NETL
NSTL
万方数据