Laplacian spectrum and coherence analysis of weighted hypercube network. (October 2018)

• Record Type: Journal Article
• Authors: Dai, Meifeng; He, Jiaojiao; Zong, Yue; Ju, Tingting; Sun, Yu; Su, Weiyi
• Is Part Of: Chaos, solitons and fractals. Volume 115(2018)
• Page Start: 29
• Page End: 34
• Publication Date: 2018-10
• Subjects: Weighted hypercube network -- Characteristic polynomial -- Laplacian spectrum -- Network coherence -- Robustness; Chaotic behavior in systems -- Periodicals; Solitons -- Periodicals; Fractals -- Periodicals; Chaotic behavior in systems; Fractals; Solitons; Periodicals; 003.7
• DOI: 10.1016/j.chaos.2018.08.012
• Languages: English
• ISSNs: 0960-0779

Abstract

Highlights: The Laplacian spectrum on the weighted hypercube network is calculated. The sum and square sum of reciprocals of all nonzero Laplacian eigenvalues is calculated. The first- and second-network coherence on weighted hypercube network is studied. The robustness of the weighted hypercube network is obtained. Abstract: Hypercube network is one of the most important and attractive network topologies so far. In this paper, we consider the scaling for first- and second-order network coherence on the hypercube network controlled by a weight factor. Our objective is to quantify the robustness of algorithms to stochastic disturbances at the nodes by using a quantity called network coherence which can be characterized as Laplacian spectrum. Network coherence can capture how well a network maintains its formation in the face of stochastic external disturbances. Firstly, we deduce the recursive relationships of its eigenvalues at two successive generations of Laplacian matrix. Then, we obtain the Laplacian spectrum of Laplacian matrix. Finally, we calculate the first- and second-order network coherence quantified as the sum and square sum of reciprocals of all nonzero Laplacian eigenvalues by using Squeeze Theorem. The obtained results show that the network coherence depends on generation number and weight factor. Meanwhile, the scalings of the first- and second-order network coherence of weighted hypercube decrease with the increasing of weight factor r, when 0 < r < 1.