Random heterogeneity outperforms design in network synchronization

The paper establishes the FS ansatz, algebraic conditions (6a)–(6b), and the delay variational/characteristic formulation (7)–(8), and then demonstrates—primarily numerically—three key claims: (i) small random heterogeneity monotonically improves stability; (ii) at intermediate σ, random heterogeneity stabilizes while designed heterogeneity (that preserves the identical orbit) does not; and (iii) for sufficiently strong disorder, the phase-locked state ceases to exist, yielding an incoherence–coherence–incoherence transition. These are clearly stated and documented with computations and experiments, but not proved rigorously in generality . The model’s solution provides a plausible analytic framework: an IFT-based existence result near σ=0, a 2×2 transfer-matrix/monodromy reduction for the directed ring characteristic exponents, perturbative formulas showing negative quadratic drift of the MTLE in case (A), and a first-order nonmonotone shift in case (B). However, crucial steps remain under-justified: (a) the IFT Jacobian invertibility is asserted via numerics rather than proved; (b) the purported uniform-in-realization open interval I on which Λ(σ)<0 for almost every realization in (A) is not established (the bound depends on the specific realization, so uniformity is nontrivial); and (c) the large-σ “pigeonhole/arctangent-closure” nonexistence argument overlooks the tan-branch/2πm bookkeeping and therefore is not rigorous. Thus, both the paper and the model offer compelling, mutually consistent narratives for the observed phenomena, but each leaves key analytic gaps.