[Paper Review] Biharmonic Distance and the Performance of Second-Order Consensus Networks with Stochastic Disturbances

Yuhao Yi, Bingjia Yang+2 more|arXiv (Cornell University)|Sep 8, 2017

Complex Network Analysis Techniques20 references4 citations

TL;DR

This paper introduces biharmonic distance as a novel metric to analyze second-order consensus networks under stochastic disturbances, showing it governs system performance similarly to how resistance distance governs first-order systems. The authors derive closed-form expressions for performance measures—pairwise variance, deviation from average, and total variance—on complete graphs, stars, cycles, and paths, and define a biharmonic Kirchhoff index and vertex centrality based on this metric.

ABSTRACT

We study second order consensus dynamics with random additive disturbances. We investigate three different performance measures: the steady-state variance of pairwise differences between vertex states, the steady-state variance of the deviation of each vertex state from the average, and the total steady-state variance of the system. We show that these performance measures are closely related to the biharmonic distance; the square of the biharmonic distance plays similar role in the system performance as resistance distances plays in the performance of first-order noisy consensus dynamics. We further define the new concepts of biharmonic Kirchhoff index and vertex centrality based on the biharmonic distance. Finally, we derive analytical results for the performance measures and concepts for complete graphs, star graphs, cycles, and paths, and we use this analysis to compare the asymptotic behavior of the steady-variance in first- and second-order systems.

Motivation & Objective

• To address the lack of a unified graph-theoretic metric analogous to resistance distance for second-order consensus systems with stochastic disturbances.
• To quantify system performance in second-order consensus networks under random additive noise using variance-based metrics.
• To establish a theoretical framework linking biharmonic distance to network coherence and system robustness in second-order dynamics.
• To define and analyze new concepts—biharmonic Kirchhoff index and biharmonic vertex centrality—for characterizing network performance and node importance.
• To compare asymptotic behavior of steady-state variance between first- and second-order consensus systems using analytical results on standard graph topologies.

Proposed method

• Proposes biharmonic distance based on the Laplacian spectrum, extending its use from computer graphics to network dynamics.
• Defines three performance measures: steady-state variance of pairwise differences, variance from average per node, and total system variance.
• Derives analytical expressions for these measures using spectral graph theory and complex analysis, particularly involving eigenvalues and eigenvectors of the Laplacian matrix.
• Applies Fourier analysis and summation identities over roots of unity to evaluate trigonometric sums arising from eigenvalue decomposition on regular graphs.
• Introduces the biharmonic Kirchhoff index as the sum of squared biharmonic distances between all node pairs.
• Defines biharmonic vertex centrality as the inverse of the sum of squared biharmonic distances from a node to all others, identifying nodes with lower steady-state variance.

Experimental results

Research questions

• RQ1 How does biharmonic distance relate to the steady-state variance in second-order consensus networks with stochastic disturbances?
• RQ2 Can a unified metric analogous to resistance distance be defined for second-order consensus systems?
• RQ3 What are the closed-form expressions for performance measures (pairwise variance, deviation variance, total variance) in standard graph topologies like complete graphs, stars, cycles, and paths?
• RQ4 How does the biharmonic Kirchhoff index and biharmonic vertex centrality compare to their first-order counterparts in terms of system coherence?
• RQ5 What is the asymptotic behavior of steady-state variance in second-order systems relative to first-order systems across different network structures?

Key findings

• The steady-state variance of pairwise differences between node states is directly governed by the square of the biharmonic distance between the nodes.
• The variance of each node's state from the system average is minimized for nodes with higher biharmonic centrality, defined as the inverse of the sum of squared biharmonic distances to all other nodes.
• For a complete graph of size $N$, the total steady-state variance is $\frac{N}{3} - \frac{1}{3N}$, derived using summation identities over trigonometric functions.
• For a cycle of size $N$, the closed-form expression for the total variance is $F_N(l) = \frac{l^4}{12N} - \frac{l^3}{3} + \frac{l^2N}{3} - \frac{l^2}{6N} + \frac{(-1)^l}{8N} + \frac{l}{3} - \frac{1}{8N}$, valid for $l \leq N$.
• The biharmonic Kirchhoff index is defined as the sum of squared biharmonic distances over all unordered node pairs, providing a global measure of network coherence in second-order systems.