Stability of Multi-Microgrids: New Certificates, Distributed Control, and Braess’s Paradox

The paper’s Theorem 1 states exactly the four claims the model proves: (a) a simple zero eigenvalue of J, (b) all nonzero real eigenvalues of J are negative, (c) under −Qi − Vi^2 Bii ≤ d_i^2/(2 m_i) all nonzero eigenvalues lie in the open left half-plane, and (d) in the lossless case all nonzero eigenvalues lie in the open left half-plane, hence local asymptotic stability (modulo uniform angle shift). These statements and hypotheses appear verbatim in the paper and are proved in Appendix B using a max-norm row argument and the identity Qk = −Vk^2 Bkk − Lkk (equations (15)–(17) leading to inequality (16)) as well as a block-matrix argument for purely imaginary eigenvalues; see Theorem 1 and Appendix B (a–d) and the derivation of (16)–(17) in the paper . The model gives a different, correct proof of (c) by showing strict row-diagonal dominance of H(λ)=λ^2M+λD+L on Re λ≥0 via a monotonicity bound on |Hii(λ)| and then applying Levy–Desplanques/Gershgorin; it also gives a concise quadratic-form argument for (d). Both approaches use the same structural facts (quadratic pencil, L’s M-matrix/Laplacian properties, and Q–B–L identity). Hence both are correct, with substantially different techniques.