Back to search
2104.14070

Foundations of Static and Dynamic Absolute Concentration Robustness (Part I of Dynamic ACR Quadrilogy)

Badal Joshi, Gheorghe Craciun

correctmedium confidenceCounterexample detected
Category
Not specified
Journal tier
Specialist/Solid
Processed
Sep 28, 2025, 12:56 AM

Audit review

The paper’s Theorem 6.1 correctly establishes (i) A1⇔A2⇔A3 and (ii) B1⇔B2⇔B3 in the complex-balanced setting, and shows Bi ⇒ Aj and full equivalence under global attraction, via the standard Horn–Jackson log-parameterization log Z = log x̃ + S⊥ and a compatibility argument that uses e_i ∈ S . By contrast, the candidate solution asserts that for mass-action systems (without the complex-balance hypothesis) dynamic ACR automatically has full basin (B1 ⇒ B3), which is false: the paper itself gives mass-action counterexamples that are dynamic ACR but narrow basin (not full) (Ex. 7) . The candidate’s proof of B1 ⇒ B3 also overlooks positivity in the compatibility step; replacing y by y + ((a*−y_i)/v_i) v need not remain in R^n_{>0}. The rest of the candidate’s arguments (A1⇔A2 via log-coset of S⊥; A2⇔A3 via an integer-lattice argument; Bi ⇒ Aj in the complex-balanced setting; equivalence under global attraction) align with the paper’s results .

Referee report (LaTeX)

\textbf{Recommendation:} reject

\textbf{Journal Tier:} specialist/solid

\textbf{Justification:}

While the candidate reproduces the main complex-balanced equivalences for static ACR and some implications among the statements, it asserts a crucially false general statement—namely, that dynamic ACR implies full basin for all mass-action systems—contradicted by explicit counterexamples in the paper. The proof step used to derive full basin overlooks positivity constraints in the compatibility construction. Given that basin distinctions are central to the paper’s dynamic ACR taxonomy, this flaw is substantive and necessitates rejection absent major revisions.