First, the concept of cooperative conflict is presented, and the characteristic of cooperative air combat is researched. Then, four methods of conflict resolution are designed by way of the first order predicate logic...First, the concept of cooperative conflict is presented, and the characteristic of cooperative air combat is researched. Then, four methods of conflict resolution are designed by way of the first order predicate logic, i.e., link-up, coordination, accommodation and integration, and corresponding examples are given. A 2 vs 2 air combat simulation was carried out; after conflict resolution, the loss ratio is dropped to 0.54 from the original 1.32, so the enhancement of effectiveness is notable. The present research findings are that the wide conflicts discover the essence of multi-fighter cooperation, i.e., to as fully as possible enhance the effectiveness of each fighter to attain global optimization, and that the possibility of conflict resolution shows the application prospect. The proposed method in this paper is a helpful try to the application of the Fifth Generation Computer in the new generation of C3I system.展开更多
常见的聚类方法存在对初始点敏感和易陷入局部最优的不足,为此提出了一种改进HBO的聚类方法。首先,提出一种改进的HBO,即扰动替换的HBO(disturbance and replacement HBO,DRHBO)克服其不足,即采用一种随机维度值替换策略和高斯扰动机制...常见的聚类方法存在对初始点敏感和易陷入局部最优的不足,为此提出了一种改进HBO的聚类方法。首先,提出一种改进的HBO,即扰动替换的HBO(disturbance and replacement HBO,DRHBO)克服其不足,即采用一种随机维度值替换策略和高斯扰动机制用于HBO中最优个体的状态更新,解决HBO搜索效率低的问题;提出一种正弦差分扰动策略,以突破当前个体仅与直接领导和同事进行交流的限制,从而增强搜索能力;将随机维度值替换和随机差分扰动策略融合,用于HBO中前期个体状态更新以避免其产生无效解。其次,提出一种DRHBO聚类方法,并运用到宫颈细胞数据集上以获得更好的聚类效果。大量、不同类别和不同样本的宫颈细胞数据集实验结果表明,与HBO及其改进算法和其他最先进算法相比,DRHBO的优化性能更好、稳定性更强且效率更高。DRHBO聚类方法更适应于宫颈细胞数据集。展开更多
基金Aeronautical Science Foundation !( 97D5 3 0 3 7)
文摘First, the concept of cooperative conflict is presented, and the characteristic of cooperative air combat is researched. Then, four methods of conflict resolution are designed by way of the first order predicate logic, i.e., link-up, coordination, accommodation and integration, and corresponding examples are given. A 2 vs 2 air combat simulation was carried out; after conflict resolution, the loss ratio is dropped to 0.54 from the original 1.32, so the enhancement of effectiveness is notable. The present research findings are that the wide conflicts discover the essence of multi-fighter cooperation, i.e., to as fully as possible enhance the effectiveness of each fighter to attain global optimization, and that the possibility of conflict resolution shows the application prospect. The proposed method in this paper is a helpful try to the application of the Fifth Generation Computer in the new generation of C3I system.
基金国家自然科学基金资助项目(12004275)Shanxi Scholarship Council of China(2020-042)山西省自然科学基金资助项目(20210302123186)。
文摘【目的】为了更精准地预测5G基站的流量,分析潮汐现象,提出一种优化的生成对抗网络(generative adversarial network,GAN)模型流量预测方法,并将其用于实际基站的定时控制中。【方法】GAN的生成器利用差分演化灰狼算法优化长短时记忆网络(long short term memory networks,LSTM),判别器使用门控循环神经网络(gated recurrent unit,GRU)进行判别,生成器和判别器利用不断地对抗训练达到均衡从而提高了5G基站流量的预测精度;其次,利用改进人工蜂群优化k-means++算法,将其用于输出最优基站定时时间,达到最大限度节能的目的。【结果】实验结果表明,与现有模型相比,所提预测模型有更高的预测精度,定时控制功能可极大地节约能耗。
文摘常见的聚类方法存在对初始点敏感和易陷入局部最优的不足,为此提出了一种改进HBO的聚类方法。首先,提出一种改进的HBO,即扰动替换的HBO(disturbance and replacement HBO,DRHBO)克服其不足,即采用一种随机维度值替换策略和高斯扰动机制用于HBO中最优个体的状态更新,解决HBO搜索效率低的问题;提出一种正弦差分扰动策略,以突破当前个体仅与直接领导和同事进行交流的限制,从而增强搜索能力;将随机维度值替换和随机差分扰动策略融合,用于HBO中前期个体状态更新以避免其产生无效解。其次,提出一种DRHBO聚类方法,并运用到宫颈细胞数据集上以获得更好的聚类效果。大量、不同类别和不同样本的宫颈细胞数据集实验结果表明,与HBO及其改进算法和其他最先进算法相比,DRHBO的优化性能更好、稳定性更强且效率更高。DRHBO聚类方法更适应于宫颈细胞数据集。
