Chen Lei, Ma Xiujuan, Ma Fuxiang, Li Yalan
College of Computer Science, Qinghai Normal University, Qinghai, 810000, China.
The State Key Laboratory of Tibetan Intelligent Information Processing and Application, Qinghai Normal University, Qinghai, 810016, China.
Sci Rep. 2024 Dec 28;14(1):31199. doi: 10.1038/s41598-024-82394-8.
The research on hypernetworks robustness focuses on improving their ability to resist various risks such as attacks and disasters. In the face of deliberate attacks, there is a huge risk of failure in Barabási-Albert (BA) hypernetworks. However, the methods to improve the risk resistant capacity of BA hypernetwork are lack. In this paper, two optimization mechanisms are proposed based on structural characteristics and connectivity. They are the self-optimization mechanism of random recombination for hyperedges(RRH) and the self-optimization mechanism of low degree preference recombination for hyperedges(LDPRH). The best improvement performance (BIP) is used as the evaluation metric, and a comparative analysis is conducted to assess the effectiveness of these mechanisms in enhancing the robustness of BA hypernetworks through simulation experiments. The research results indicate that both mechanisms effectively improve the hypernetwork's risk resistance, but the optimization effects differ. When the number of recombined hyperedges is 0.2*M (where M is the total number of hyperedges in the hypernetwork), the BIP of the random recombination mechanism stabilize at 25%, while the BIP of the low preference recombination mechanism ultimately reaches about 65%. To further validate the effectiveness of these optimization mechanisms, they were applied to Newman-Watts (NW) hypernetworks and Barabási-Albert (BA) ordinary networks, the latter based on binary relationships. The results showed that these optimization mechanisms also improved the risk resistance of NW hypernetworks and BA ordinary networks, especially in the BA ordinary network, where the BIP could reach up to 73%. Additionally, a significant correlation is observed between BIP, the number of hyperedge reshuffling operations, and the uniformity of the hypernetwork. Moreover, these optimization mechanisms are further validated in the context of China's high-speed railway hypernetwork, highlighting their practical application potential.
超网络鲁棒性的研究主要集中在提高其抵御诸如攻击和灾难等各种风险的能力。面对蓄意攻击，巴拉巴西-阿尔伯特（BA）超网络存在巨大的失败风险。然而，目前缺乏提高BA超网络抗风险能力的方法。本文基于结构特征和连通性提出了两种优化机制。它们分别是超边随机重组的自优化机制（RRH）和超边低度偏好重组的自优化机制（LDPRH）。以最佳改进性能（BIP）作为评估指标，通过仿真实验进行对比分析，以评估这些机制在增强BA超网络鲁棒性方面的有效性。研究结果表明，这两种机制都能有效提高超网络的抗风险能力，但优化效果有所不同。当重组超边的数量为0.2*M（其中M是超网络中超边的总数）时，随机重组机制的BIP稳定在25%，而低度偏好重组机制的BIP最终达到约65%。为了进一步验证这些优化机制的有效性，将它们应用于纽曼-瓦特（NW）超网络和基于二元关系的巴拉巴西-阿尔伯特（BA）普通网络。结果表明，这些优化机制也提高了NW超网络和BA普通网络的抗风险能力，特别是在BA普通网络中，BIP可达73%。此外，还观察到BIP、超边重排操作的数量与超网络的均匀性之间存在显著相关性。此外，这些优化机制在中国高速铁路超网络的背景下得到了进一步验证，突出了它们的实际应用潜力。
