A counterexample to the Han–Wang–Yu spectral hull conjecture
Abstract
Gluing two stochastic matrices yields an eigenvalue outside their spectral hull.
Definition 1.1 (Finite chains with a selected root).
Formalization. D5/S3/Quantum/Matrix/HanWangYuGluingSpectrumRefutation.RootedChain (✓ std3).
Citation. X. B. Han; S. Wang; C. Yu (2025). PageRank of Gluing Networks and Corresponding Markov Chains. DOI: 10.3390/math13132080. URL: https://www.mdpi.com/2227-7390/13/13/2080.
Commentary.
Definition 8, page 10: “For integers i = 1, 2, ⋯, k, let M_i = (S_i, A_i) be k Markov chains, and v_0^i be a fixed point in S_i. Assume θ = (θ_1, θ_2, ⋯, θ_k), 0 < θ_i < 1, and Σ_{i=1}^k θ_i = 1.” RootedChain has fields n, A and stochastic. Its states are Fin (n + 1), with selected root 0. The real matrix A is column stochastic: entry (v,w) is the transition probability from w to v. Source indices 1 through k become Fin k.
Definition 1.2 (The transition matrix after gluing).
Formalization. D5/S3/Quantum/Matrix/HanWangYuGluingSpectrumRefutation.glue (✓ std3).
Citation. X. B. Han; S. Wang; C. Yu (2025). PageRank of Gluing Networks and Corresponding Markov Chains. DOI: 10.3390/math13132080. URL: https://www.mdpi.com/2227-7390/13/13/2080.
Commentary.
Definition 8, page 10: “The θ-gluing of M_i along v_0^i is the following Markov chain M = (S, A).” The four cases give the source transition rule. The common root is none, and some ⟨i,v⟩ is state v.succ of input i. The transport h ▸ v changes the private index from chain i to chain j when h : i = j.
Definition 1.3 (The proposed spectral containment).
Formalization. D5/S3/Quantum/Matrix/HanWangYuGluingSpectrumRefutation.claim (✓ std3).
Citation. X. B. Han; S. Wang; C. Yu (2025). PageRank of Gluing Networks and Corresponding Markov Chains. DOI: 10.3390/math13132080. URL: https://www.mdpi.com/2227-7390/13/13/2080.
Commentary.
Conjecture 1, §4.3, page 17: “Glue Markov chains M_i = (S_i, A_i) to a Markov chain M = (S, A), then all the eigenvalues of A lie in the convex hull of the eigenvalues of all A_i on the complex plane.” The spectrum is taken over ℂ after applying algebraMap ℝ ℂ entrywise. The real convex hull lies in the complex plane. Every weight is strictly between zero and one and the weights sum to one.
Theorem 1.4 (The spectral containment fails).
Proof. Machine-checked in Lean as D5/S3/Quantum/Matrix/HanWangYuGluingSpectrumRefutation.result (✓ std3). ∎
Resolves. Problems/han-wang-yu-gluing-spectrum (refuted) by D5/S3/Quantum/Matrix/HanWangYuGluingSpectrumRefutation.result.
Source. Repository-derived.
Acknowledgement. X. B. Han; S. Wang; C. Yu (2025). PageRank of Gluing Networks and Corresponding Markov Chains. DOI: 10.3390/math13132080. URL: https://www.mdpi.com/2227-7390/13/13/2080.
Commentary.
Take two copies of B = (1/54)[[43,9,2],[2,36,16],[9,9,36]], rooted at 0, with weights (1/2,1/2). Every column sums to one and every entry is nonnegative. The characteristic polynomial is (μ−1)(μ−17/27)(μ−1/2), so every input spectral value has real part at least 1/2. The glued vector (0,−4,3,4,−3) has eigenvalue 4/9: the private blocks have opposite signs and their contributions to the root row cancel. Convexity of the half-plane Re z ≥ 1/2 excludes 4/9 from the input spectral hull.
References
- Truth anchor:
D5/S3/Quantum/Matrix/HanWangYuGluingSpectrumRefutation.RootedChain - Truth anchor:
D5/S3/Quantum/Matrix/HanWangYuGluingSpectrumRefutation.claim - Truth anchor:
D5/S3/Quantum/Matrix/HanWangYuGluingSpectrumRefutation.glue - Truth anchor:
D5/S3/Quantum/Matrix/HanWangYuGluingSpectrumRefutation.result - Dependency: D5/S3/Combinatorics/NonBacktracking/LaplacianGapMinDegree