Bounds on eigenvalue ratios of quantum graphs

Abstract: We study ratios of eigenvalues of the Laplacian on compact metric graphs. Our goals are threefold: First, we prove a sharp Ashbaugh--Benguria-type bound for the ratio of the first two eigenvalues on compact trees with Dirichlet conditions at all leaves, concretely showing that the ratio is maximized when the graph is an interval or an equilateral star. This improves a previous Payne--Pólya--Weinberger-type result due to Nicaise [Bull. Sci. Math., II. Sér. 111 (1987), 401--413]. Second, we extend this bound to a set of inequalities for the ratio of any pair of eigenvalues of such compact Dirichlet trees which respect the Weyl asymptotics up to an absolute constant. Third, we show that on non-trees, on which we also allow any mix of Neumann and Dirichlet conditions at the leaves, it is possible to recover bounds on the eigenvalue ratios depending only on the number of independent cycles and the number of Neumann leaves, in addition to the eigenvalue indices. This complements previously known counterexamples to analogues of the Ashbaugh--Benguria bound for general quantum graphs, by showing that the only way the bound can fail is through cycles and Neumann leaves, and by explicitly quantifying the extent to which it can fail.

• The paper establishes sharp bounds on eigenvalue ratios for quantum graphs, proving that for Dirichlet trees, λ2/λ1 is at most 4 and extending these results to general eigenvalue pairs.
• It employs innovative techniques such as edge length perturbations and surgical arguments to optimize eigenvalue ratios, connecting nodal domain counts with spectral bounds.
• The analysis clearly shows how cycles and Neumann leaves modify eigenvalue behavior by introducing explicit dependence in the bounds, linking classical inequalities with Weyl asymptotics.

Sharp Bounds on Eigenvalue Ratios for Quantum Graph Laplacians

Introduction and Context

This paper ["Bounds on eigenvalue ratios of quantum graphs" (2603.26172)] addresses the problem of estimating ratios of Laplacian eigenvalues for compact metric graphs, specifically Laplacians with mixes of Dirichlet, Neumann (Kirchhoff), and delta-type boundary conditions. The problem is motivated by classical inequalities for Euclidean domains (such as Ashbaugh–Benguria and Payne–Polya–Weinberger bounds), but the paper adapts and significantly sharpens these results for the case of quantum graphs, which introduce topological and boundary-condition-induced complexities absent in Euclidean domains.

The study focuses on three main goals: (1) establishing a sharp Ashbaugh–Benguria-type bound for the first two eigenvalues in Dirichlet trees, (2) extending such bounds to general eigenvalue ratios for Dirichlet trees, and (3) quantifying the failure of absolute bounds in non-tree graphs with cycles and Neumann leaves, resulting in bounds dependent on cycle number and boundary configuration.

Main Results: Sharp Bounds on Dirichlet Trees

For compact Dirichlet trees—graphs with no cycles and Dirichlet boundary conditions imposed at all leaves—the authors prove that the ratio of the second to the first eigenvalue of the Laplacian satisfies

$$\frac{\lambda_2 (\Gamma)}{\lambda_1 (\Gamma)} \leq 4$$

Equality is attained precisely for intervals (single-edge graphs) or equilateral star graphs, generalizing the 1D ball/interval case and identifying a broader family of extremal graphs. The proof utilizes surgical arguments and optimization with respect to edge length perturbation, departing from traditional symmetrization techniques.

Furthermore, this sharp ratio bound is extended to arbitrary pairs of eigenvalues. For all Dirichlet trees and all $k>j\geq 1$:

$$\frac{\lambda_k (\Gamma)}{\lambda_j (\Gamma)} \leq 4 \cdot \frac{k^2}{j^2}$$

This result is significant because it matches the leading order Weyl asymptotics (for the Laplacian on an interval), up to the absolute constant 4, and provides a substantial improvement over exponentially growing previous bounds (e.g., $5^{k-j}$ from Nicaise's classical result).

Structural Drivers of Bound Failure for General Quantum Graphs

For general compact graphs—including those with cycles and mixed Neumann/Dirichlet boundary conditions—the paper demonstrates that universal absolute upper bounds for eigenvalue ratios can fail. The only obstructions to sharp universal bounds are independent cycles (first Betti number $\beta$) and Neumann leaves ($N$). For $k > j \geq N+\beta+1$, the eigenvalue ratio admits the following explicit upper bound:

$$\frac{\lambda_k (\Gamma)}{\lambda_j (\Gamma)} \leq 4\cdot \frac{k^2}{(j-(N+\beta))^2}$$

This bound precisely quantifies the deviation induced by cycles and Neumann leaves, offering a complete answer as to why counterexamples arise for Ashbaugh–Benguria-type inequalities in the quantum graph setting.

Optimization and Existence: Extremizers among Graph Classes

The paper further establishes the existence of both minimizers and maximizers for eigenvalue ratios within given topological classes (e.g., graphs with specified numbers of edges, cycles, Dirichlet, and Neumann leaves). These extremizers always exist due to the continuity of eigenvalues with respect to edge length perturbations (Hadamard-type formulas), and may involve degenerate cases with edges of zero length, corresponding to graph contraction.

For Dirichlet trees, the infimum of $\lambda_k/\lambda_j$ is always 1, achieved by appropriate equilateral star graphs with Dirichlet conditions at the center. The maximization problem, especially for higher-index ratios, is left as an open question, with strong numerical and heuristic evidence suggesting that intervals maximize ratios for $\lambda_k/\lambda_1$, at least for small $k>j\geq 1$0.

Analytical Techniques and Implications

The key technical innovation is the use of edge length perturbations and surgery principles to optimize eigenvalue ratios, combined with precise variational characterizations. For trees, the authors exploit nodal domain counts: after generic perturbations, the $k>j\geq 1$1-th eigenfunction in a Dirichlet tree graph has precisely $k>j\geq 1$2 nodal domains, allowing clean reduction to subgraph eigenvalue comparisons and application of sharp bounds.

The results rigorously identify the interplay between graph topology and boundary conditions in spectral optimization, connecting the structure of quantum graphs to both classical inequalities and Weyl-type asymptotics.

Practical and Theoretical Implications

Practically, these results enable tight estimates for spectral gaps and eigenvalue ratios on quantum graphs, which are critical in quantum chaos, photonics, network analysis, and analysis of vibration/transport properties in discrete-continuum models. The explicit dependence on cycle number and boundary configuration informs design strategies for engineered networks, offering quantitative spectral control.

Theoretically, the sharp bounds and structural characterization present a natural quantum graph analogue of classical domain results, providing conjectural guidance for spectral optimization in graph-based models—both for mathematical physics and spectral geometry. The identification of extremal configurations also links to combinatorial optimization and inverse spectral problems.

Future Directions

Several open problems remain, such as determining the optimal constant (whether $k>j\geq 1$3 or possibly $k>j\geq 1$4) in general eigenvalue ratio bounds for Dirichlet trees, and characterizing equality cases in Pólya-type inequalities. Further work could extend these optimization principles to broader classes of graphs (with restricted numbers of leaves and cycles) and explore spectral extremizers outside the tree case.

The foundational techniques, particularly the edge perturbation and surgery frameworks, are expected to be relevant for future spectral optimization and extremal eigenvalue problems in quantum graph theory and related discrete-continuum systems.

Conclusion

The paper delivers sharp, explicit, and nearly complete bounds for eigenvalue ratios of Laplacians on quantum graphs, precisely delineating the interaction between graph topology and vertex boundary conditions. The results bridge classical spectral inequalities and quantum graph analysis, and provide powerful tools for spectral optimization in both theoretical and applied contexts.