Pith. sign in

REVIEW 3 major objections 6 minor 39 references

Distributed Averaging Problems over Directed Signed Networks

T0 review · 3 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper shows that signed-average consensus on strongly connected directed signed networks is achievable exactly when the signed digraph is structurally balanced—and state stability exactly when it is unbalanced—without requiring weight…

desk verdict Useful cofactor-weighted Laplacian construction for signed-average consensus on directed graphs, but Theorem 2's proof needs a short repair. read the letter →

arxiv 1908.06250 v2 pith:U5OBL3X4 submitted 2019-08-17 math.OC math.DS

Mingjun Du , Deyuan Meng , Zheng-Guang Wu This is my paper · ORCID
classification math.OCmath.DS MSC 93A1693D5005C50
keywords signednetworkssigned-averageconsensusdistributedaveragingstructuralbalancedirecteddigraphsLaplacianpotentialfixed-timemulti-agentsystems
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

In a signed network, agents cooperate with some neighbors and compete with others, and the goal of distributed averaging is for every agent to end at the same magnitude, with signs prescribed by the network's camp structure. Earlier solutions required the digraph to be weight balanced, meaning each agent's total incoming influence had to equal its outgoing influence. This paper removes that restriction for strongly connected directed topologies by introducing an improved Laplacian potential, built by weighting each edge with a cofactor determinant of the induced unsigned Laplacian and symmetrizing the graph into an undirected mirror signed graph. Using either of two proposed nearest-neighbor protocols, the paper proves signed-average consensus if and only if the signed digraph is structurally balanced, and state stability, meaning all states converge to zero, if and only if it is structurally unbalanced. The same potential also yields fixed-time bipartite consensus or stability, with settling-time bounds independent of initial conditions.

What carries the argument

The central object is the improved Laplacian potential function $\Phi_e(x) = \sum_{i,j} \det(L_{ii}) |a_{ij}| (x_i - \operatorname{sgn}(a_{ij}) x_j)^2 = x^T (W L + L^T W) x$, where $W$ is the diagonal matrix of cofactors $\det(L_{ii})$ of the induced unsigned Laplacian. For a strongly connected digraph these cofactors are strictly positive and, as a vector, form a left eigenvector of $L$ for eigenvalue zero; this is what makes the mirror Laplacian $\hat L = (W L + L^T W)/2$ symmetric with the same null-space structure as $L$. The potential does the work of a Lyapunov function: along $\dot x = -\hat L x$, the derivative of $V(x) = x^T x$ is exactly $-\Phi_e(x)$, so LaSalle's invariance principle turns the null-space dichotomy of $\hat L$ into the consensus-versus-stability dichotomy, with no weight-balance condition anywhere. The same potential supports the Lyapunov estimate for the fixed-time nonlinear protocol in Theorem 3.

What would settle it

Run protocol (12) on a strongly connected, digon sign-symmetric signed digraph that is structurally balanced but has unequal in- and out-degrees, starting from random initial states; the paper predicts the terminal state is exactly $\frac{1}{n}(1_n^T D_n x_0) D_n 1_n$. If the computed terminal vector differs beyond numerical error, or if $\hat L = (W L + L^T W)/2$ has a zero eigenvector not proportional to $D_n 1_n$, the core dichotomy fails. Repeating the test with a reciprocal edge pair whose weights have opposite signs, violating digon sign-symmetry, locates the boundary of the assumption.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that weight imbalance was never the real obstruction to signed-average consensus on directed signed networks; the sign pattern is. For a strongly connected signed digraph $G$, take absolute values of the edge weights to form the induced unsigned digraph, let $W = \mathrm{diag}(\det(L_{11}), \dots, \det(L_{nn}))$ be the diagonal matrix of principal-minor cofactors of its Laplacian, and define the mirror signed graph by the symmetrized Laplacian $\hat L = (W L + L^T W)/2$. Because the vector of cofactors is a left eigenvector of $L$ for eigenvalue $0$, $\hat L$ is a genuine symmetric Laplacian, and its null space is $\mathrm{span}\{D_n 1_n\}$ exactly when $G$ is structurally balanced and $\{0\}$ otherwise. The systems $\dot x = -\hat L x$ (protocol (9)) and $\dot x = -W L x$ (protocol (12)) therefore converge, respectively, to the signed average $(1/n)\sum_j \sigma_j x_{j0}$ applied with the gauge signs, or to zero. Theorems 1 and 2 state these as if-and-only-if results: structural balance is necessary as well as sufficient, and structural unbalance is necessary as well as sufficient for stability.

Load-bearing premise

The load-bearing assumption is that the signed digraph is digon sign-symmetric—for every pair of agents, the product of the two reciprocal edge weights is nonnegative—so that no two agents influence each other with opposite signs; the mirror graph inherits the sign pattern only under this condition, and the if-and-only-if consensus and stability theorems depend on that inheritance.

Editorial extensions

If this is right

  • Signed-average consensus becomes attainable on every strongly connected directed signed network, weight-balanced or not, and the limiting value is the gauge-transformed average of the initial states.
  • The two linear protocols (9) and (12) give simple implementations: one is the gradient of the improved potential, the other applies cofactor gains to the classical neighbor rule.
  • The balanced/unbalanced dichotomy is sharp: structurally unbalanced digraphs drive all states to zero rather than to a polarized consensus, so the sign pattern alone decides the outcome.
  • With the nonlinear protocol (18), the same conclusions hold in fixed time, with explicit upper bounds on the settling time that do not depend on initial states.
  • Because unsigned networks are signed networks with all-positive edges, both protocols also solve average consensus on strongly connected unsigned digraphs without weight balance.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural next step, not pursued here, is relaxing digon sign-symmetry; the mirror-graph sign-preservation lemma is where the assumption enters, so mixed-sign reciprocal edge pairs are the first place to look for counterexamples.
  • The cofactor gains used in the protocols are principal minors of the global Laplacian, so although the update rules are local, the protocols are not fully distributed in information; estimating these gains online or replacing them with local surrogates would make the design practically distributed.
  • The symmetrization recipe—weighting a directed Laplacian by a positive diagonal matrix from its left null space before symmetrizing—does not depend on the averaging problem and could transfer other directed-network results, such as synchronization, formation control, and opinion dynamics, to weight-unbalanced digraphs.
  • A quantitative prediction is that the convergence rate of the mirror-Laplacian protocol is governed by the spectral gap of the mirror graph; changing edge weights while keeping sign patterns would test whether weight imbalance affects only the rate, not the outcome.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper studies distributed averaging over signed directed networks. It introduces two graph-theoretic constructions—the induced unsigned digraph and the mirror (undirected) signed graph—and uses cofactors of the induced unsigned Laplacian to define an "improved Laplacian potential" that is a quadratic form in the signed Laplacian even when the digraph is not weight balanced. Two linear nearest-neighbor protocols are proposed, one based on the mirror Laplacian and one on the cofactor-weighted signed Laplacian, and the paper claims that signed-average consensus (respectively, state stability) is achieved if and only if the strongly connected signed digraph is structurally balanced (respectively, unbalanced). A nonlinear protocol is then claimed to achieve fixed-time bipartite consensus or state stability with a settling-time bound independent of initial conditions. The simulations illustrate the claims. All results are explicitly conditional on a digon sign-symmetry assumption stated in Section III-A.

Significance. If the theorems are correct, the paper makes a valuable contribution by removing the weight-balance restriction for signed-average consensus on directed signed networks and by providing a Lyapunov-based convergence analysis that works for weight-unbalanced topologies. The fixed-time extension is a useful application of the improved potential. The paper builds transparently on established tools (the cofactor property from [37] and the fixed-time lemmas from [30]) rather than assuming its conclusions. However, the proof of Theorem 2 contains a false intermediate assertion, and the proof of Lemma 3 has a gap; these are repairable, but as written they compromise the validity of the main if-and-only-if claims.

major comments (3)
  1. [Theorem 2 proof, Section IV (system (15) and Eq. (3))] The proof asserts '[det(L11),...,det(Lnn)]L = 0_n^T' for the signed Laplacian L appearing in (15), citing [37]. The cofactor vector is a left null vector of the induced unsigned Laplacian L̄ = Δ − |A|, not of the signed Laplacian L_s = Δ − A. For the digon sign-symmetric, strongly connected, structurally balanced digraph with adjacency A = [[0,1,-1],[1,0,-1],[-1,-1,0]], one computes L_s = [[2,-1,1],[-1,2,1],[1,1,2]], W = 3I_3, and 1_3^T W L_s = [6,6,12] ≠ 0. Thus the claimed equality is false, and the subsequent identification of ν_l = D_3 1_3 as a left eigenvector of WL_s is not established by the published proof. The conclusion is salvageable: from L_s = D L̄ D and w^T L̄ = 0 one obtains (D1)^T W L_s = (Dw)^T L̄ D = 0. This step must be written out, and the signed-average value formula then follows by normalizing the left eigenvector correctly. As it stands, the balanced direction of Theorem 2 lacks a valid proof.
  2. [Lemma 3, Appendix C] In the converse direction of Lemma 3(1), the paper asserts that from N(L̂) = N(L) = span{D_n 1_n} one can derive D_n L D_n 1_n = L 1_n = 0_n, and that this implies D_n L D_n = L. Neither step is valid for a signed Laplacian: L 1_n is not generally zero (it equals the vector of signed row sums), and a single null vector does not determine the matrix. The structural-balance conclusion does follow from N(L) = span{D_n 1_n} via known results for signed Laplacians, but the argument as written does not establish it. Since Lemma 3 underpins Lemma 4 and therefore the 'only if' halves of Theorems 1 and 2, this gap needs to be repaired.
  3. [Section III-A vs. Abstract and Introduction] The paper assumes throughout that the signed digraph is digon sign-symmetric (a_ij a_ji ≥ 0 for all i,j), yet the abstract promises 'general directed topologies' and the introduction claims 'arbitrary directed signed networks.' This assumption is essential: without it, the mirror signed graph in Lemma 2 need not preserve the sign pattern, and the equivalence of structural balance between G and Ĝ can fail. For example, a two-node graph with a_12 = 1 and a_21 = −1 has cofactors det(L̄_11) = det(L̄_22) = 1, giving â_12 = (1·1 + 1·(−1))/2 = 0, so the mirror graph is disconnected and Lemma 2(1) fails. The paper should either qualify its claims to 'strongly connected, digon sign-symmetric signed digraphs' or prominently state this restriction as a limitation of the proposed approach.
minor comments (6)
  1. [Throughout] The symbol L is used for both the signed Laplacian and the induced unsigned Laplacian; Definition 1 and Eq. (3) are ambiguous. The proof error in Theorem 2 is a direct consequence of this notational conflation. Please use L_s and L̄ consistently.
  2. [Theorem 3, Eq. (19)] The settling-time formula (19) appears to have a typesetting error: the term involving k_1 should be n^{(m-r)/(2r)}/k_1^{r/(m-r)}, as follows from Lemma 7 with α_1 = k_1 n^{(r-m)/(2r)} λ_2(L̂), rather than n^{(m-r)/(2r)} k_1^{r/(m-r)}. Please verify the displayed formula.
  3. [Lemma 2] The statement that Ĝ 'has the same sign pattern as G' is imprecise: Ĝ is undirected and may contain an edge where only one directed edge exists in G, and the relevant claim is about the sign of any nonzero â_ij. Consider rewording for clarity.
  4. [Table I] Table I reports only the settling-time upper bounds; the eigenvalues λ_1(L̂) and λ_2(L̂) used in (19)-(20) should be reported so that the numerical values can be reproduced.
  5. [Proof of Theorem 3] The derivation of (23) from (22) invokes [30, Lemma 4] but does not state it; please state the lemma or make the application explicit, since it is essential to the fixed-time bound.
  6. [Editorial] There are several typos and notational slips, e.g., 'Lapalcian' in Section III-A, and the repeated use of 'L = Δ − A' with different meanings in Definition 1 and the surrounding text. A careful notation and proofreading pass is needed.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the improved-Laplacian construction and consensus results do not assume their conclusions; self-citations are to standard, parameter-free prior theorems.

full rationale

The paper's central construction is an explicit algebraic potential using cofactor weights det(Lii) taken from the induced unsigned Laplacian, not from the target consensus value or from the signed Laplacian's nullspace. Lemma 4 derives the quadratic-form identity Phi_e(x)=x^T(WL+L^TW)x from the external cofactor identity of [37] (w^T Lbar = 0); this is an independent algebraic fact, not a restatement of signed-average consensus. The if-and-only-if statements in Theorems 1 and 2 are established by Lyapunov arguments and nullspace characterizations (Lemmas 1-3), where the balancing gauge D appears as an object to be found, not as an input defining the protocol. No parameter is fitted to data and then renamed a prediction; protocol gains and cofactors are determined directly by the graph. The self-citations [26] and [30] supply eigenvalue characterizations and comparison lemmas that are parameter-free external theorems with stated hypotheses (strong connectivity, signed digraph properties); they do not smuggle in structural balance or signed-average consensus. A separate concern that the proof of Theorem 2 appears to conflate the induced unsigned Laplacian Lbar with the signed Laplacian L when asserting 1^T W L = 0 is a proof-correctness issue rather than circularity, and does not change the circularity verdict.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The central claim rests on standard graph-theoretic and Lyapunov tools, plus the digon sign-symmetry and strong connectivity domain assumptions. No free parameters are fitted to data, and no new physical or conceptual entities are postulated. The main caveat is that the cofactor weights det(Lbar_ii) are global graph information, which is an implementation assumption not listed as an axiom.

assumptions (5)
  • domain assumption Every considered signed digraph is digon sign-symmetric, aij * aji >= 0.
    Stated in Section III-A. Lemma 2's sign-preservation property of the mirror graph and the validity of the mirror Laplacian rely on this assumption.
  • domain assumption The signed digraph is strongly connected.
    Used to guarantee positive cofactors, the null-space characterizations in Lemma 3, and the convergence results in Theorems 1, 2, and 3.
  • standard math For a strongly connected unsigned digraph, all principal minors det(Lbar_ii) are positive and the cofactor vector is a positive left null vector of Lbar.
    Taken from [37] and used to define W and derive the balancing identity (27), which is central to the improved Laplacian potential.
  • standard math LaSalle's invariance principle and the comparison lemma for differential inequalities.
    Used in the proofs of Theorems 1 and 3 to establish convergence and settling-time bounds.
  • standard math Lemma 7 from [30] on fixed-time convergence of a scalar differential equation.
    Used in Theorem 3 to bound the settling time after reducing the Lyapunov analysis to a scalar comparison system.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Distributed Averaging Problems over Directed Signed Networks." pith.science (2026). https://pith.science/paper/U5OBL3X4

@misc{pith2026190806250,
  author       = {Pith},
  title        = {Pith review of: Distributed Averaging Problems over Directed Signed Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/U5OBL3X4}},
  note         = {Machine review of arXiv:1908.06250}
}
read the original abstract

This paper aims at addressing distributed averaging problems for signed networks in the presence of general directed topologies that are represented by signed digraphs. A new class of improved Laplacian potential functions is proposed by presenting two notions of any signed digraph: induced unsigned digraph and mirror (undirected) signed graph, based on which two distributed averaging protocols are designed using the nearest neighbor rules. It is shown that with any of the designed protocols, signed-average consensus (respectively, state stability) can be achieved if and only if the associated signed digraph of signed network is structurally balanced (respectively, unbalanced), regardless of whether weight balance is satisfied or not. Further, improved Laplacian potential functions can be exploited to solve fixed-time consensus problems of signed networks with directed topologies, in which a nonlinear distributed protocol is proposed to ensure the bipartite consensus or state stability within a fixed time. Additionally, the convergence analyses of directed signed networks can be implemented with the Lyapunov stability analysis method, which is realized by revealing the tight relationship between convergence behaviors of directed signed networks and properties of improved Laplacian potential functions. Illustrative examples are presented to demonstrate the validity of our theoretical results for directed signed networks.

Figures

Figures reproduced from arXiv: 1908.06250 by the authors.

Figure 1
Figure 1. Two signed digraphs Ga and Gb. Left: Ga is structurally balanced. Right: Gb is structurally unbalanced. Example 1. The initial states of agents are provided by x(0) = [1, 2, 3, 4, 5, 6]T . Consider the signed network (1) under the signed digraph Ga in [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 3
Figure 3. Signed-average consensus of the signed network (1). Upper: Under the protocol (9). Lower: Under the protocol (12). are plotted in [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 2
Figure 2. Bipartite consensus of the signed network (1) by employing the protocol (13). By using the protocol (13), we can plot the state evolution of the signed network (1) in [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: State stability of the signed network (1) by employing the protocol (13). . 0 1 2 3 4 5 −1 0 1 2 3 4 5 6 Time t Agent State xi x1 x2 x3 x4 x5 x6 0 1 2 3 4 5 −4 −3 −2 −1 0 1 2 3 4 5 6 Time t Agent State xi x1 x2 x3 x4 x5 x6 [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 6
Figure 6. Figure 6: Fixed-time bipartite consensus under the structurally balanced signed digraph Ga. Upper: x(0) = [2, 2, 3, 4, 3, 6]T . Lower: x(0) = [5, −10, 15, −5, 10, −15]T . VII. CONCLUSIONS In this paper, we have investigated the distributed averaging problems of signed networks w…
Figure 5
Figure 5. Figure 5: State stability of the signed network (1). Upper: Under the protocol (9). Lower: Under the protocol (12). TABLE I: The setting time T with different initial states T(≤) [2, 2, 3, 4, 3, 6]T [5, −10, 15, −5, 10, −15]T Ga 0.5850 0.5850 Gb 2.1835 2.1835 network (1) employi…
Figure 7
Figure 7. Figure 7: Fixed-time state stability under the structurally unbalanced signed digraph Gb. Upper: x(0) = [2, 2, 3, 4, 3, 6]T . Lower: x(0) = [5, −10, 15, −5, 10, −15]T . APPENDIX A. Proof of Lemma 1 Proof: Let Lˆ and ∆ = diag ˆ {∆ˆ 11, ∆ˆ 22, · · · , ∆ˆ nn} denote the Laplacian m…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 38 canonical work pages

  1. [37]

    Algebraic criteria for consensus prob lems of continuous-time network systems,

    Z. Li and Y . Jia, “Algebraic criteria for consensus prob lems of continuous-time network systems,” International Journal of Control, vol. 82, no. 4, pp. 643-658, 2009

    work page 2009

  2. [30]

    Finite-time consensus for mu ltiagent sys- tems with cooperative and antagonistic interactions,

    D. Meng, Y . Jia, and J. Du, “Finite-time consensus for mu ltiagent sys- tems with cooperative and antagonistic interactions,” IEEE Transactions on Neural Networks and Learning Systems , vol. 27, no. 4, pp. 762-770, Apr. 2016. 12

    work page 2016

  3. [1]

    Consensus seeking in multi-agent sy stems under dynmicaly changing interaction topologies,

    W. Ren and R. Beard, “Consensus seeking in multi-agent sy stems under dynmicaly changing interaction topologies,” IEEE Transactions on Automatic Control , vol. 50, no. 5, pp. 655-661, May 2005

    work page 2005

  4. [2]

    Reaching a cons ensus in a dynamically changing environment: A graphical approac h,

    M. Cao, A. S. Morse, and B. D. O. Anderson, “Reaching a cons ensus in a dynamically changing environment: A graphical approac h,” SIAM Journal on Control and Optimization , vol. 47, no. 2, pp. 575-600, 2008

    work page 2008

  5. [3]

    Some necessary and sufficient c on- ditions for second-order consensus in multi-agent dynamic al systems,

    W. Y u, G. Chen, and M. Cao, “Some necessary and sufficient c on- ditions for second-order consensus in multi-agent dynamic al systems,” Automatica, vol. 46, pp. 1089-1095, 2010

    work page 2010

  6. [4]

    Distributed control of ac tive-passive net- worked multiagent systems,

    T. Y ucelen and J. D. Peterson, “Distributed control of ac tive-passive net- worked multiagent systems,” IEEE Transactions on Control of Network Systems, vol. 4, no. 4, pp. 707-715, Dec. 2017

    work page 2017

  7. [5]

    Distributed consensus al gorithms for a class of high-order multi-agent systems on directed graph s,

    A. Abdessameud and A. Tayebi, “Distributed consensus al gorithms for a class of high-order multi-agent systems on directed graph s,” IEEE Transactions on Automatic Control , vol. 63, no. 10, pp. 3464-3470, Oct. 2018

    work page 2018

  8. [6]

    Finite-time connectivit y-preserving consensus for second-order nonlinear multiagent systems,

    H. Hong, W. Y u, J. Fu, and X. Y u, “Finite-time connectivit y-preserving consensus for second-order nonlinear multiagent systems, ” IEEE Trans- actions on Control of Network Systems , vol. 6, no. 1, pp. 236-248, Mar. 2019

    work page 2019

Show all 39 references
  1. [7]

    Output consensus of hete rogeneous linear multi-agent systems with adaptive event-triggered control,

    Y .-Y . Qian, L. Liu, and G. Feng, “Output consensus of hete rogeneous linear multi-agent systems with adaptive event-triggered control,” IEEE Transactions on Automatic Control , vol. 64, no. 6, pp. 2606-2613, Jun. 2019

    2019

  2. [8]

    Adaptive protocol design for distributed tracking with relative output information: A distributed fi xed-time ob- server approach,

    Y . Lv, G. Wen, and T. Huang, “Adaptive protocol design for distributed tracking with relative output information: A distributed fi xed-time ob- server approach,” IEEE Transactions on Control of Network Systems , vol. 7, no. 1, pp. 118-128, Mar. 2020

    2020

  3. [9]

    Consensus problems in networks of agents with switching topology and time-delays,

    R. Olfati-Saber and R. M. Murray, “Consensus problems in networks of agents with switching topology and time-delays,” IEEE Transactions on Automatic Control , vol. 49, no. 9, pp. 1520-1533, Sept. 2004

    2004

  4. [10]

    Average consensus on general stron gly connected digraphs,

    K. Cai and H. Ishii, “Average consensus on general stron gly connected digraphs,” Automatica, vol. 48, no. 11, pp. 2750-2761, Nov. 2012

    2012

  5. [11]

    A d istributed algo- rithm for average consensus on strongly connected weighted digraphs,

    A. Priolo, A. Gasparri, E. Montijano, and C. Sagues, “A d istributed algo- rithm for average consensus on strongly connected weighted digraphs,” Automatica, vol. 50, no. 3, pp. 946-951, Mar. 2014

    2014

  6. [12]

    Average consensus in networ ks of dynamic multi-agents with switching topology: Infinite matrix prod ucts,

    H. Atrianfar and M. Haeri, “Average consensus in networ ks of dynamic multi-agents with switching topology: Infinite matrix prod ucts,” ISA Transactions, vol. 51, no. 4, pp. 522-530, Jul. 2012

    2012

  7. [13]

    Average consensus on arbitrary str ongly connected digraphs with time-varying topologies,

    K. Cai and H. Ishii, “Average consensus on arbitrary str ongly connected digraphs with time-varying topologies,” IEEE Transactions on Automatic Control, vol. 59, no. 4, pp. 1066-1071, Apr. 2014

    2014

  8. [14]

    Average consen sus in the presence of delays in directed graph topologies,

    C. N. Hadjicostis and T. Charalambous, “Average consen sus in the presence of delays in directed graph topologies,” IEEE Transactions on Automatic Control , vol. 59, no. 3, pp. 763-768, Mar. 2014

    2014

  9. [15]

    Weighted average conse nsus-based unscented kalman filtering,

    W. Li, G. Wei, F. Han, and Y . Liu, “Weighted average conse nsus-based unscented kalman filtering,” IEEE Transactions on Cybernetics , vol. 46, no. 2, pp. 558-567, Feb. 2016

    2016

  10. [16]

    A tutorial on modeling and analysis of dynamic social networks. Part II,

    A. V . Proskurnikov and R. Tempo, “A tutorial on modeling and analysis of dynamic social networks. Part II,” Annual Reviews in Control , vol. 45, pp. 166-190, 2018

    2018

  11. [17]

    Consensus problems on networks with antag onistic inter- actions,

    C. Altafini, “Consensus problems on networks with antag onistic inter- actions,” IEEE Transactions on Automatic Control , vol. 58, no. 4, pp. 935-946, Apr. 2013

    2013

  12. [18]

    Bipartite consensus of multi-age nt systems over signed graphs: State feedback and output feedback control a pproaches,

    H. Zhang and J. Chen, “Bipartite consensus of multi-age nt systems over signed graphs: State feedback and output feedback control a pproaches,” International Journal of Robust and Nonlinear Control , vol. 27, pp. 3- 14, Jan. 2017

    2017

  13. [19]

    On the bipartite co nsensus for generic linear multiagent systems with input saturatio n,

    J. Qin, W. Fu, W. X. Zheng, and H. Gao, “On the bipartite co nsensus for generic linear multiagent systems with input saturatio n,” IEEE Transactions on Cybernetics , vol. 47, no. 8, pp. 1948-1958, Aug. 2017

    1948

  14. [20]

    Consensus control of general linear multi-agent systems with antagonistic interaction s and communi- cation noises,

    J. Hu, Y . Wu, T. Li, and B. K. Ghosh, “Consensus control of general linear multi-agent systems with antagonistic interaction s and communi- cation noises,” IEEE Transactions on Automatic Control , vol. 64, no. 5, pp. 2122-2127, May 2019

    2019

  15. [21]

    G. Wen, H. Wang, X. Y u, and W. Y u, Bipartite tracking cons ensus of linear multi-agent systems with a dynamic leader, IEEE Transactions on Circuits and System–II: Express Briefs , vol. 65, no. 9, pp. 1024-1028, Sept. 2018

    2018

  16. [22]

    Prescribed performance for b ipartite tracking control of nonlinear multiagent systems with hyst eresis input uncertainties,

    T. Y u, L. Ma, and H. Zhong, “Prescribed performance for b ipartite tracking control of nonlinear multiagent systems with hyst eresis input uncertainties,” IEEE Transactions on Cybernetics , vol. 49, no. 4, pp. 1327-1338, Apr. 2019

    2019

  17. [23]

    Biparti te tracking of homogeneous and heterogeneous linear multi-agent syste ms,

    Q. Jiao, H. Zhang, S. Xu, F. L. Lewis, and L. Xie, “Biparti te tracking of homogeneous and heterogeneous linear multi-agent syste ms,” Inter- national Journal of Control , vol. 92, no. 12, pp. 2963-2972, 2019

    2019

  18. [24]

    Interval bipartite consensu s of networked agents associated with signed digraphs,

    D. Meng, M. Du, and Y . Jia, “Interval bipartite consensu s of networked agents associated with signed digraphs,” IEEE Transactions on Auto- matic Control, vol. 61, no. 12, pp. 3755-3770, Dec. 2016

    2016

  19. [25]

    Structural balance and opin- ion sparation in trust-mistrust social networks

    W. Xia, M. Cao, and K. H. Johansson, “Structural balance and opin- ion sparation in trust-mistrust social networks” IEEE Transactions on Control of Network Systems , vol. 3, no. 1, pp. 46-56, Mar. 2016

    2016

  20. [26]

    Convergence analysis of directed signed netw orks via an M matrix approach,

    D. Meng, “Convergence analysis of directed signed netw orks via an M matrix approach,” International Journal of Control , vol. 91, no. 4, pp. 827-847, 2018

    2018

  21. [27]

    Beh aviors of Networks with Antagonistic Interactions and Switching T opologies,

    Z. Meng, G. Shi, K. H. Johansson, M. Cao, and Y . Hong, “Beh aviors of Networks with Antagonistic Interactions and Switching T opologies,” Automatica, vol. 73, pp. 110-116, Nov. 2016

    2016

  22. [28]

    Opinion d ynamics in social networks with hostile camps: Consensus vs. polariza tion,

    A. V . Proskurnikov, A. S. Matveev, and M. Cao, “Opinion d ynamics in social networks with hostile camps: Consensus vs. polariza tion,” IEEE Transactions on Automatic Control , vol. 61, no. 6, pp. 1524-1536, Jun. 2016

    2016

  23. [29]

    Swarming behav ior of multiple euler-lagrange systems with cooperation-comp etition inter- actions: An auxiliary system approach,

    H. Hu, G. Wen, W. Y u, Q. Xuan, and G. Chen, “Swarming behav ior of multiple euler-lagrange systems with cooperation-comp etition inter- actions: An auxiliary system approach,” IEEE Transactions on Neural Networks and Learning Systems , vol. 29, no. 11, pp. 5726-5737, Nov. 2018

    2018

  24. [31]

    Signed-average consensus for netwo rks of agents: A nonlinear fixed-time convergence protocol,

    D. Meng and Z. Zuo, “Signed-average consensus for netwo rks of agents: A nonlinear fixed-time convergence protocol,” Nonlinear Dynamics, vol. 85, no. 1, pp. 155-165, Jul. 2016

    2016

  25. [32]

    Adaptive finite-time biparti te consensus for second-order multi-agent systems with antagonistic in teractions,

    L. Zhao, Y . Jia, and J. Y u, “Adaptive finite-time biparti te consensus for second-order multi-agent systems with antagonistic in teractions,” Systems and Control Letters , vol. 102, pp. 22-31, Apr. 2017

    2017

  26. [33]

    Finite-time biparti te consensus for multi-agent systems on directed signed networks,

    H. Wang, W. Y u, G. Wen, and G. Chen, “Finite-time biparti te consensus for multi-agent systems on directed signed networks,” IEEE Transac- tions on Circuits and System–I: Regular Papers , vol. 65, no. 12, pp. 4336-4348, Dec. 2018

    2018

  27. [34]

    Finite-time and fixed-time bi partite consensus of multi-agent systems under a unified discontinu ous control protocol,

    X. Liu, J. Cao, and C. Xie, “Finite-time and fixed-time bi partite consensus of multi-agent systems under a unified discontinu ous control protocol,” Journal of Franklin Institute , vol. 356, no. 2, pp. 734-751, Jan. 2019

    2019

  28. [35]

    Finite-time bipartit e consensus for multiagent systems under detail-balanced antagonisti c interactions,

    J. Lu, Y . Wang, X. Shi, and J. Cao, “Finite-time bipartit e consensus for multiagent systems under detail-balanced antagonisti c interactions,” IEEE Transactions on Systems, Man, and Cybernetics: System s, to be published, DOI: 10.1109/TSMC.2019.2938419

  29. [36]

    Robust consensus algorithms for mul tiscale coordination control of multivehicle systems with disturb ances,

    D. Meng and Y . Jia, “Robust consensus algorithms for mul tiscale coordination control of multivehicle systems with disturb ances,” IEEE Transactions on Industrial Electronics , vol. 63, no. 2, pp. 1107-1119, Feb. 2016

    2016

  30. [38]

    H. K. Khalil, Nonlinear Systems , NJ, USA: Prentice Hall, 2002

    2002

  31. [39]

    Consistency in marked digr aphs,

    L. W. Beineke and F. Harary, “Consistency in marked digr aphs,” Journal of Mathematical Psychology , vol. 18, no. 3, pp. 260-269, Dec. 1978

    1978

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.