Imitation dynamics in population games on community networks

The paper proves (i) that in potential population games over undirected, connected community networks the population state approaches the set Y◦, via a Lyapunov–LaSalle argument (Lemma 2) and a boundary-exclusion Lemma 1, culminating in Theorem 2; and (ii) under fully supported isolated Nash equilibria and Assumption 2, convergence to a single Nash equilibrium and to the balanced state y*η^> (Corollary 2, with Proposition 3 for the x-limit) . The candidate reproduces the Lyapunov part and the aggregate dynamics ẏ with Λij(x) = (XWX^T)ij correctly (matching (24)–(25)) , and the balanced-consensus limit for x at y* (aligned with Proposition 3) . However, the candidate’s proof has two critical gaps: (a) it infers limt dist(y(t), Y◦) = 0 from showing Φ̇ ≥ 0 and limt dist(y(t), Y•) = 0, without the paper’s additional Lemma 1 needed to exclude boundary restricted equilibria not in Y◦; and (b) it incorrectly claims that full support along the trajectory forces accumulation points to have full support, which need not hold. These errors mean the model’s proof does not justify the paper’s main convergence claim to Y◦ (Theorem 2) nor the reduction to Nash equilibria without the extra lemma/assumptions, even though the end statements match the paper’s results. The paper’s arguments are internally consistent and complete (Assumptions 1–2; Proposition 1, Lemma 2, Theorem 2, Proposition 3, Corollary 2) .