REVIEW 3 major objections 5 minor 44 references
Sensor nodes on a Boolean control network can track a global state average with a proven almost-sure error bound using only local observations and neighbor exchanges.
Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →
T0 review · grok-4.5
2026-07-31 16:03 UTC pith:XJWDOS4N
load-bearing objection Solid distributed filter for stochastic BCNs with an explicit a.s. practical-consensus bound; the math is coherent, the ergodicity assumption is the real soft spot, and it deserves referees. the 3 major comments →
Stochastic Average Consensus Filtering and Distributed State Estimation for Boolean Control Networks
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
Under conditional independence of sensor observations, geometric ergodicity of the closed-loop Boolean network, and a mild step-size bound, the distributed stochastic average consensus filter drives every node’s estimate of the network-wide average posterior to a compact invariant set whose radius is governed by algebraic connectivity and sensor heterogeneity; the limsup of the consensus error is therefore finite almost surely and can be reduced by strengthening the graph.
What carries the argument
A Radon–Nikodym change of measure that converts each local observation model into an independent process, yielding an STP-based recursive formula for the local conditional state expectation; those expectations are then fused by a fixed-gain stochastic-approximation consensus iteration whose almost-sure practical convergence is established via a perturbed stochastic Lyapunov function and the martingale convergence theorem.
Load-bearing premise
The closed-loop Boolean network under the chosen state-feedback law must be geometrically ergodic; without that mixing property the stationary mean field and the vanishing-perturbation argument collapse.
What would settle it
Run the algorithm on a connected five-node sensor graph with the paper’s step-size rule and a non-ergodic feedback gain; if the measured limsup of the consensus error still stays inside the predicted c1 ball (or fails to diverge), the geometric-ergodicity hypothesis is unnecessary; if it exceeds c1 under an ergodic gain that satisfies all stated assumptions, the bound itself is false.
If this is right
- Global state estimation for large Boolean networks becomes possible without a fusion center or raw observation exchange.
- Raising the algebraic connectivity of the sensor graph systematically shrinks the almost-sure error ball.
- When the averaged posterior satisfies the separation condition, every node recovers the same unique MAP state; otherwise the admissible candidate set is explicitly characterized.
- The same local-filter-plus-consensus template applies to any sensor subset chosen by an observer, from one node up to the full network.
Where Pith is reading between the lines
- The same change-of-measure plus stochastic-approximation pattern should extend, with only local modifications, to Boolean networks whose communication graph switches among a finite family of connected topologies.
- Because the error radius is monotone in λ2, topology co-design (adding a few high-impact edges) can be used as an explicit tuning knob to meet a prescribed estimation accuracy before any filtering runs.
- If packet dropouts are modeled as random edge erasures that preserve connectivity in expectation, the same Lyapunov argument is likely to yield a larger but still finite almost-sure bound.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper develops a distributed multi-sensor state estimation scheme for disturbed Boolean control networks. Local conditional state expectations are computed recursively via a change-of-measure argument and the semi-tensor product (Theorem 2). These heterogeneous local beliefs are then fused by a fixed-step-size stochastic-approximation consensus iteration (20) that uses only neighbor exchanges. Under conditional independence of sensors, geometric ergodicity of the closed-loop BCN, and a step-size restriction, the consensus error is shown to be almost surely ultimately bounded by an explicit constant c_1 that depends on algebraic connectivity and sensor heterogeneity (Proposition 2, Theorem 5); a MAP extraction rule with a separation condition is given in Theorem 6. A five-sensor numerical example illustrates bounded consensus error and MAP recovery.
Significance. Distributed estimation for logical/BCN systems is genuinely less developed than its continuous-state counterpart, and the combination of STP-based Bayesian recursion with stochastic-approximation consensus is a natural and useful contribution. The explicit dependence of the error radius on λ_2(L) is attractive for network design. The appendices supply detailed martingale and perturbed-Lyapunov arguments in the style of Kushner–Yin, which is the right technical toolkit. If the gaps around ergodicity and the stationary mean-field characterization are closed, the result would be a solid reference point for distributed filtering on discrete logical networks.
major comments (3)
- [Assumption 2; §5 (matrix Q); Theorem 5] Assumption 2 (geometric ergodicity of {x(k)} under the designed SFC u=Kx) is load-bearing for Proposition 1, Lemma 5, and therefore Theorem 5: it supplies the unique stationary law, the SLLN for the mean field, and vanishing cumulative perturbations. In §5 a concrete K and the closed-loop matrix Q=FKΦ_3 are exhibited, yet irreducibility/aperiodicity (hence geometric ergodicity on a finite set) of Q are never checked. The almost-sure limsup claim is conditional on a property that is assumed rather than verified for the running example. Please either prove that the designed Q is irreducible and aperiodic, or replace Assumption 2 by a verifiable structural condition on (F,K) and confirm it in the simulation.
- [Proposition 1; Appendix D] Proposition 1 asserts that the stationary mean ē^s of the local filter satisfies the deterministic fixed-point equation obtained by substituting the marginal y^s(∞) into the nonlinear Bayesian update (19). Because of the Hadamard product and the normalization, expectation does not pass inside the update: the mean of the stationary filter belief is not equal to the filter applied to the mean observation. The existence of ē=lim E[e_k] still follows from geometric ergodicity of the filter (cf. the cited Le Gland–Mevel theory), so the ODE mean field Λη+Γ(ē−ē_avg) remains well-defined, but the explicit fixed-point formula is incorrect and should be removed or replaced by a correct characterization (e.g., via the stationary measure on the belief simplex).
- [Proposition 2, Eq. (30); Lemma 3] In Proposition 2 the radius is written c_1=√((1+d_max)^2−2λ_2(L))·∥ē−ē_avg∥·|λ_max(Λ)|^{-1}. The radicand can become negative when algebraic connectivity is large relative to d_max; the Lyapunov argument then actually yields a stricter bound (possibly c_1=0 when α=0), but the formula as stated is not always real. Please give a well-defined expression (e.g., replace the square root by the operator norm bound ∥Γα∥ that is always valid) and clarify the regular-graph versus non-regular-graph case, since Λ1=−(1+d) is not a scalar multiple of 1 on irregular graphs and the consensus subspace is not invariant without the careful cancellation you use later.
minor comments (5)
- [Abstract; §4.2] Notation switches between P-a.s. and almost-sure practical convergence; state once that the limsup bound is the precise claim and use it consistently in the abstract and §4.2.
- [§4.1, Eq. (20); Lemma 3] Equation (20) and Lemma 3 use |N| for the number of sensors while N=2^n is already reserved for the size of the state space; rename the sensor count (e.g., r) everywhere to avoid clash.
- [§5] Figures 1–7 are referenced with only brief captions; add axis labels, units, and a one-sentence takeaway to each caption so the simulation section is self-contained.
- [References] Several bibliography entries are incomplete or inconsistently formatted (e.g., [25], [26], [29], [42]); normalize to the journal style and fix missing venues/pages.
- [Throughout] Typographical issues: “Index T erms”, “Ap×q-dimensional”, repeated “the the”, and minor spacing errors around STP symbols. A careful copy-edit pass is needed.
Circularity Check
No circularity: error bound and consensus claims are derived from SA/ODE analysis under stated assumptions, not fitted or self-defined from the target.
full rationale
The load-bearing chain is: local posterior recursion via measure change and STP (Theorems 1–2, Appendices A–C); distributed SA update (20); mean-field ODE η̇=Λη+Γ(ē−ē_avg) under geometric ergodicity (Assumption 2, Prop. 1); Lyapunov c1-stability of that ODE (Prop. 2); Kushner–Yin tracking plus martingale/perturbed-Lyapunov recurrence to get limsup ∥εk∥≤c1 a.s. (Thm. 5, Lemmas 4–5, Apps. D–G). The constant c1 is an explicit function of graph eigenvalues, dmax, and stationary local-estimate heterogeneity—not calibrated to simulation accuracy or defined in terms of the limsup it bounds. Simulations illustrate trajectories and step-size effects; they do not fit parameters that are then re-presented as predictions. Citations (STP, HMM filtering, Kushner–Yin, prior consensus filters) are external tools; no load-bearing uniqueness theorem or ansatz is imported from overlapping-author work that would force the main claim by construction. Assumption 2 is an unverified modeling hypothesis for the designed K, which is a correctness/scope risk, not circularity. Derivation is self-contained against its own equations and standard external theorems.
Axiom & Free-Parameter Ledger
free parameters (2)
- step size ρ =
example values 0.05–0.3 in Fig. 7; no universal optimum claimed
- graph edge weights a_ij =
explicit 5×5 A given in simulation
axioms (6)
- domain assumption Assumption 1: sensor observations are mutually conditionally independent given x(k).
- domain assumption Assumption 2: {x(k)} is geometrically ergodic under the state-feedback u=Kx.
- ad hoc to paper Assumption 3: 0<ρ<2(1+3d_max)^{-1}.
- domain assumption Communication graph G is undirected, simple, and connected (so λ_2(L)>0).
- standard math Semi-tensor product algebra and algebraic state-space representation of BCNs (Cheng et al.).
- standard math Kushner–Yin / Benveniste stochastic-approximation ODE method and martingale convergence theorems.
read the original abstract
This paper addresses the distributed multi-sensor fusion state estimation and consensus filtering for Boolean control networks (BCNs). Existing centralized multi-sensor estimation schemes for stochastic BCNs have limitations of high communication costs and single-point failures, and continuous-state consensus algorithms are difficult to extend to discrete logical systems. By integrating probability measure transformation, semi-tensor product and stochastic approximation, a distributed stochastic average consensus filter is proposed. Moreover, the almost sure convergence of the algorithm is proved by martingale convergence theorem and perturbed stochastic Lyapunov functions. The proposed framework realizes global state estimation via local communication, avoiding the defects of centralized architectures.
Figures
Reference graph
Works this paper leans on
-
[1]
S. A. Kauffman, ”Metabolic stability and epigenesis in randomly constructed genetic nets,”J. Theor. Biol., vol. 22, no. 3, pp. 437–467, Mar. 1969
1969
-
[2]
Cheng, H
D. Cheng, H. Qi, and Z. Li, ”Analysis and control of Boolean networks,”IEEE Trans. Neural Netw. Learn. Syst., vol. 25, no. 10, pp. 1854–1866, Oct. 2014
2014
-
[3]
Cheng and H
D. Cheng and H. Qi, ”Semi-tensor product of matrices and its applications,”Sci. China Inf. Sci., vol. 52, no. 1, pp. 1–38, Jan. 2009
2009
-
[4]
H. Qi, D. Cheng, and Z. Li, ”Stability and stabilization of Boolean networks,”Int. J. Robust Nonlinear Control, vol. 20, no. 11, pp. 1213–1225, Jul. 2010
2010
-
[5]
Cheng, H
D. Cheng, H. Qi, and Y. Lin, ”Controllability and observability of Boolean control networks,”Automatica, vol. 45, no. 7, pp. 1659–1667, Jul. 2009
2009
-
[6]
Z. Li, D. Cheng, and H. Qi, ”Optimal control of Boolean control networks,”IEEE Trans. Autom. Control, vol. 56, no. 4, pp. 876–882, Apr. 2011
2011
-
[7]
Y. Liu, J. Lu, and J. Cao, ”Fault detection for Boolean networks via semi-tensor product,”IEEE Trans. Cybern., vol. 45, no. 12, pp. 2839–2848, Dec. 2015
2015
-
[8]
Wang, ”Stochasticity in gene expression: From theories to phenotypes,”Nat
J. Wang, ”Stochasticity in gene expression: From theories to phenotypes,”Nat. Rev. Genet., vol. 10, no. 12, pp. 865–876, Dec. 2009. 21
2009
-
[9]
M. B. Elowitz, A. J. Levine, E. D. Siggia, and P. S. Swain, ”Stochastic gene expression in a single cell,”Science, vol. 297, no. 5584, pp. 1183–1186, Aug. 2002
2002
-
[10]
Shmulevich, E
I. Shmulevich, E. R. Dougherty, S. Kim, and W. Zhang, ”Probabilistic Boolean networks: A rule-based uncertainty model for gene regulatory networks,”Bioinformatics, vol. 18, no. 2, pp. 261–274, Feb. 2002
2002
-
[11]
Zhang, J
H. Zhang, J. Lam, and X. Mao, ”Stability analysis of Markov jump Boolean networks,”IEEE Trans. Neural Netw. Learn. Syst., vol. 28, no. 11, pp. 2602–2615, Nov. 2017
2017
-
[12]
Albert and S
R. Albert and S. A. Kauffman, ”Boolean dynamics of networks with scale-free topology,”Phys. Rev. Lett., vol. 86, no. 21, pp. 4993–4996, May 2001
2001
-
[13]
M. C. O. Lima, L. S. Silva, and R. V. M. Silva, ”p53-MDM2 feedback loop: A key pathway in cancer,”Biomed. Res. Int., vol. 2014, pp. 1–10, 2014
2014
-
[14]
Bar-Shalom, X
Y. Bar-Shalom, X. R. Li, and T. Kirubarajan, ”Estimation with applications to tracking and navigation,”Wiley- Intersci., Hoboken, NJ, USA, 2001
2001
-
[15]
Haykin, ”Cognitive radar: A way of the future,”IEEE Signal Process
S. Haykin, ”Cognitive radar: A way of the future,”IEEE Signal Process. Mag., vol. 23, no. 1, pp. 30–40, Jan. 2006
2006
-
[16]
A. H. Sayed, A. Tarighat, and N. Khajehnouri, ”Network-based wireless sensor fusion for target tracking,”IEEE Signal Process. Mag., vol. 24, no. 3, pp. 61–70, May 2007
2007
-
[17]
J. N. Ash and A. H. Sayed, ”Distributed fusion for target tracking in sensor networks,”IEEE Trans. Signal Process., vol. 57, no. 11, pp. 4566–4581, Nov. 2009
2009
-
[18]
X. Wang, H. Zhang, and J. Liu, ”Multi-sensor fusion for wearable health monitoring systems,”IEEE J. Biomed. Health Inform., vol. 22, no. 6, pp. 1621–1631, Nov. 2018
2018
-
[19]
Li and Y
F. Li and Y. Tang, ”Multi-sensor fusion Boolean Bayesian filtering for stochastic Boolean networks,”IEEE Trans. Neural Netw. Learn. Syst., vol. 33, no. 10, pp. 5644–5655, Oct. 2022
2022
-
[20]
X. Li, J. Lu, and J. Cao, ”Bayesian filtering for stochastic Boolean networks,”Automatica, vol. 72, pp. 252–258, Oct. 2016
2016
-
[21]
Y. Liu, J. Lu, and J. Cao, ”Maximum likelihood estimation for Boolean networks with incomplete measurements,” IEEE Trans. Cybern., vol. 46, no. 7, pp. 1604–1615, Jul. 2016
2016
-
[22]
Wang and H
Y. Wang and H. Wu, ”Hidden Markov Boolean control networks under shifting attacks: Estimation and control,” IEEE Trans. Autom. Control, vol. 66, no. 11, pp. 5267–5282, Nov. 2021
2021
-
[23]
M. G. Rabbat and R. D. Nowak, ”Distributed optimization in sensor networks,” inProc. 3rd Int. Symp. Inf. Process. Sensor Netw., Berkeley, CA, USA, 2004, pp. 20–27
2004
-
[24]
Ghasemi, H
A. Ghasemi, H. R. Rabiee, and A. H. Sayed, ”Distributed HMM filtering using consensus-based methods,”IEEE Trans. Signal Process., vol. 60, no. 11, pp. 5847–5861, Nov. 2012
2012
-
[25]
Stochastic average consensus filter for distributed hmm filtering: Almost sure con- vergence[J]
Ghasemi N, Dey S, Baras J S. Stochastic average consensus filter for distributed hmm filtering: Almost sure con- vergence[J]. IF AC Proceedings Volumes, 2010, 43(19): 335-340
2010
-
[26]
Multi-sensor fusion particle filtering for Boolean networks with multi-step randomly-delayed mea- surements[J]
Shao S, Xiang L. Multi-sensor fusion particle filtering for Boolean networks with multi-step randomly-delayed mea- surements[J]. Neurocomputing, 2023, 547: 126386
2023
-
[27]
Asymptotic agreement in distributed estimation,
V. Borkar and P. P. Varaiya, “Asymptotic agreement in distributed estimation,”IEEE Transactions on Automatic Control, vol. AC-27, no. 3, pp. 650–655, June 1982
1982
-
[28]
Shi D , Elliott R J , Chen T .On Finite-State Stochastic Modeling and Secure Estimation of Cyber-Physical Sys- tems[J].IEEE Transactions on Automatic Control, 2016, 62(1):65-80.DOI:10.1109/TAC.2016.2541919
arXiv 2016
-
[29]
Wang L , Wu Z G .Shifting Attack Stabilization and Estimation of Hidden Markov Boolean Networks[J].IEEE Transactions on Cybernetics, 2025:1-11.DOI:10.1109/tcyb.2025.3535929
arXiv 2025
-
[30]
Consensus filters for sensor networks and distributed sensor fusion,
R. Olfati-Saber and J. S. Shamma, “Consensus filters for sensor networks and distributed sensor fusion,” inProceed- ings of the 44th IEEE Conference on Decision and Control, and the European Control Conference, Seville, Spain, December 12-15 2005, pp. 6698–6703. 22
2005
-
[31]
Distributed Kalman filtering using weighted averaging,
P. Arliksson and A. Rantzer, “Distributed Kalman filtering using weighted averaging,” inProceedings of the 17th International Symposium on Mathematical Theory of Networks and Systems, Kyoto, Japan, July 2006
2006
-
[32]
Distributed Kalman filtering based on consensus strategies,
R. Carli, A. Chiuso, L. Schenato, and S. Zampieri, “Distributed Kalman filtering based on consensus strategies,” IEEE Journal on Selected Areas in Communications, vol. 26, no. 4, pp. 622–633, May 2008
2008
-
[33]
Approximate distributed Kalman filtering in sensors networks with quantifiable performance,
D. P. Spanos, R. Olfati-Saber, and R. M. Murray, “Approximate distributed Kalman filtering in sensors networks with quantifiable performance,” inProceedings of the 4th International Symposium on Information Processing in Sensor Networks, Los Angeles, CA, USA, April 2005, pp. 133–139
2005
-
[34]
A scheme for robust distributed sensor fusion based on average consensus,
L. Xiao, S. Boyd, and S. Lall, “A scheme for robust distributed sensor fusion based on average consensus,” in Proceedings of the fourth International Symposium on Information Processing in Sensor Networks, April 2005, pp. 63–70
2005
-
[35]
Distributed information filtering using consensus filters,
D. W. Casbeer and R. Beard, “Distributed information filtering using consensus filters,” inProceedings of the 2009 American Control Conference (ACC 2009), Hyatt Regency Riverfront, St. Louis, MO, USA, June 2009, pp. 1882– 1887
2009
-
[36]
Benveniste, M
A. Benveniste, M. M´ etivier, and P. Priouret,Adaptive Algorithms and Stochastic Approximations, A. V. Balakrish- nan, I. Karatzas, and M. Yor, Eds., vol. 22, in Applications of Mathematics, Berlin Heidelberg: Springer-Verlag, 1990
1990
-
[37]
H. J. Kushner and G. G. Yin,Stochastic approximation and recursive algorithms and applications, 2nd ed., B. Ro- zovskii and M. Yor, Eds., vol. 35, in Applications of Mathematics, New York, NY, USA: Springer-Verlag New York, Inc., 2003
2003
-
[38]
Exponential Forgetting and Geometric Ergodicity in Hidden Markov Models,
F. Le Gland and L. Mevel, “Exponential Forgetting and Geometric Ergodicity in Hidden Markov Models,”Mathe- matics of Control, Signals, and Systems, vol. 13, pp. 63–93, 2000
2000
-
[39]
A. N. Shiryaev,Probability, 2nd ed., S. Axler, F. Gehring, and P. Halmos, Eds., vol. 95, in Graduate Texts in Mathematics, New York, NY, USA: Springer-Verlag New York, Inc., 1996
1996
-
[40]
Billingsley,Probability and Measure, 3rd ed
P. Billingsley,Probability and Measure, 3rd ed. New York, NY, USA: John Wiley & Sons, Inc., 1995
1995
-
[41]
Hidden Markov models: estimation and control[M]
Elliott R J, Moore J B, Aggoun L. Hidden Markov models: estimation and control[M]. New York, NY: Springer New York, 1995
1995
-
[42]
Optimal State Estimation of Boolean Control Networks under Stochastic Function Pertur- bations[J]
Li L, Guo Y, Lu J, et al. Optimal State Estimation of Boolean Control Networks under Stochastic Function Pertur- bations[J]. SIAM Journal on Control and Optimization, 2026, 64(1): 450-471
2026
-
[43]
Distributed asynchronous deterministic and stochastic gradient optimization algorithms,
J. N. Tsitsiklis, D. P. Bertsekas, and M. Athans, “Distributed asynchronous deterministic and stochastic gradient optimization algorithms,”IEEE Transactions on Automatic Control, vol. 31, no. 9, pp. 803–812, September 1986
1986
-
[44]
Distributed Kalman Filter with Embedded Consensus Filters,
R. Olfati-Saber, “Distributed Kalman Filter with Embedded Consensus Filters,” inProceedings of the 44th IEEE Conference on Decision and Control, and the European Control Conference, Seville, Spain, December 12-15 2005, pp. 8179–8184. 23
2005
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.