---
title: Brualdi–Solheid Problem in Spectral Graphs
url: https://www.emergentmind.com/topics/brualdi-solheid-problem
type: topic
---

# Brualdi–Solheid Problem in Spectral Graphs

The Brualdi–Solheid Problem refers to a class of extremal spectral graph theory questions concerning the maximum or minimum possible spectral radius of graphs within a specified family, typically subject to discrete structural constraints such as order, size, degree, forbidden subgraphs, or other invariants. This family of problems generalizes classical extremal graph theory (Turán-type) to the spectral realm, and has become central in connecting combinatorial and linear algebraic aspects of graph structure.

## 1. Problem Formulation and Fundamental Concepts

For a family of graphs $\mathscr{H}$ defined by a property (e.g., being $F$-free, fixed order/size, degree constraints), the Brualdi–Solheid Problem asks:

\[
\max\{\rho(G): G \in \mathscr{H}\}
\]

where $\rho(G)$ denotes the spectral radius (i.e., the largest eigenvalue) of the adjacency matrix of $G$; in some variants, one considers other eigenvalue functions, such as the signless Laplacian or the general $A_\alpha$-index.

Key definitions include:

- **$F$-free graph**: $G$ is $F$-free if it contains no subgraph isomorphic to $F$.
- **Extremal graph**: A graph achieving the maximum (or minimum) spectral radius in $\mathscr{H}$.
- **Classical Turán function** (edge extremal): $\mathrm{ex}(n, F) = \max\{e(G): |G|=n,\, G \text{ is } F \text{-free}\}$
- **Brualdi–Solheid Function** ("spectral Turán"): $\mathrm{spex}(n, F) = \max\{\rho(G): |G|=n,\, G \text{ is } F \text{-free}\}$

The Brualdi–Solheid problem also appears in variants maximizing $\rho(G)$ subject to fixed order and size, degree constraints, connectivity requirements, or forbidden minors [2302.03313][2511.06643][2405.20056][2412.16019].

## 2. Classical Cases and Methodological Reductions

The original instance was posed for the class $\mathcal{H}_{n,m}$: all connected simple graphs with $n$ vertices and $m$ edges. In this setting, extremal graphs maximizing $\rho(G)$ are always threshold graphs, a class defined by a recursive construction or an explicit vertex-ordering with associated generating binary sequence (Thm., [2412.16019]).

For forbidden subgraphs, the paradigm is parallel to classical Turán theory: maximizing $\rho(G)$ over all $n$-vertex $F$-free graphs. Extensively, solutions show that, in many cases, the extremal graphs are the same as in the edge extremal (Turán) setting; i.e., for large $n$ the same $r$-partite Turán graphs or their explicit joins [2305.03952][2302.03313][2110.11345][2007.06922].

For generalizations such as the $A_\alpha$-matrix, the same classification for $A(G)$-extremals persists for $A_\alpha(G)$ across small- or moderate-size regimes, with threshold graphs again playing the key role [2511.06643].

A common methodological insight is the reduction to monotone degree sequences and structural transformations (edge/vertex swaps) that strictly increase $\rho(G)$ unless the underlying structure is stabilized—this forces extremality conditions and reduces complexity, especially in connected/threshold cases [2412.16019][2511.06643].

## 3. Forbidden Subgraph Formulations: Spectral Turán Problems

In the forbidden subgraph context, the Brualdi–Solheid problem asks for the spectral counterpart to classical Turán-type results:

- For $F=C_\ell^2$, the square of the $\ell$-cycle with $\ell\ge6$ and $\ell\not\equiv0\pmod3$, the extremal graph for both maximal edge count and maximal spectral radius is $T_{n,3}$ (balanced complete 3-partite graph) for $\ell\equiv1\pmod3$, and the join $K_1+T_{n-1,3}$ for $\ell\equiv2\pmod3$ [2305.03952].
- For even cycles $C_{2k+2}$ and trees on $2k+3$ vertices in bipartite graphs, the unique extremal graphs are the complete bipartite $K_{k,n-k}$ [2302.03313].
- For consecutive odd cycles $C_{2\ell+1}$ up to a fixed length in non-bipartite graphs, the extremal structure is $K_{a, b}\bullet K_{3}$, the complete bipartite $K_{a,b}$ with a triangle "grafted" via vertex identification [2110.11345].

The spectral extremal problem for wheel-free graphs—graphs not containing any $W_k$ for $k \geq 4$—has also been completely resolved, characterizing the extremal family $H_n$ depending on $n \pmod{4}$ and, for the signless Laplacian, a distinct join structure $K_2 \nabla (n-2)K_1$ [2007.06922].

A recurring theme is that, for most standard forbidden subgraphs (cycles, trees, wheels, and certain bipartite structures), the extremal graphs for the spectral radius mirror those for extremal edge count, with only exceptional configurations or parity restrictions yielding distinct extremals.

## 4. Spectral Extremal Results for Trees and Generalized Parameters

For the family of all graphs with given matching number or vertex/edge-connectivity properties, the Brualdi–Solheid problem admits both minimization and maximization forms:

- **Graphs with given matching number $\beta$**: The minimizer for the spectral radius over $\mathcal{G}_{n,\beta}$ is always a tree, structurally classified via the novel *quasi-adjacency* relation that encodes clique block-decomposition among control sets (Theorem 2.6 etc., [2601.18223]). Extremal trees are constructed by balancing the sizes of attached stars and central paths; explicit formulas are provided for $\beta=2,3,4$.
- **Graphs with prescribed $h$-extra $r$-component connectivity**: For $\mathcal{G}_{n,\delta}^{\kappa_r^h}$, extremal graphs maximizing $\rho(G)$ are unique, characterized by a join of cliques with minimal spread of missing edges to achieve the required degree and cut conditions [2405.20056].
- **Connected threshold graphs with fixed order and size**: Brualdi–Solheid extremals are always threshold graphs, and the paper provides tight spectral upper and lower bounds in terms of parameters of the binary generating sequence and the structure of "lazy walks" [2412.16019].

## 5. Core Techniques and Major Theoretical Insights

- **Eigenvalue and Rayleigh Quotient Argumentation**: Critical for establishing extremality under spectral optimization, especially when coupled with monotonicity properties under local modifications (edge addition/move increases $\rho(G)$ unless structural balance is achieved) [2511.06643][2405.20056].
- **Structural Reductions to Threshold Graphs**: In fixed $(n,m)$ connected settings, extremality is only possible for threshold graphs, which admit a recursive construction and simple degree ordering [2412.16019][2511.06643].
- **Spectral Stability and Perturbation Lemmas**: For forbidden subgraph problems, small deviations from extremality in the spectral radius imply $o(n^2)$-edge differences from the conjectured extremal template, enforceable via Perron vector analysis and edge-swapping [2305.03952].
- **Combinatorial–Algebraic Interplay**: Explicit use of eigenvalue interlacing, vertex partitioning by Perron vector entries, and local combinatorial arguments (path/cycle embeddings, component decompositions) to translate global spectral constraints into strong structural characterizations [2302.03313][2007.06922][2110.11345].
- **Characteristic Polynomial Engineering**: In the matching-minimal case, recurrence relations and block-gluing formulas for characteristic polynomials yield precise control over the possible spectral radii for candidate trees [2601.18223].

## 6. Open Problems and Directions

- The Brualdi–Solheid maximal spectral radius problem for all $n, m$ and for certain families of forbidden subgraphs (especially for intermediate or large $m$) remains unresolved and may require refined local-global structural techniques beyond threshold graph reductions [2511.06643][2412.16019].
- For $A_\alpha$-spectral radius with general $\alpha \in [0,1)$, characterizations are complete only for substantial ranges of $m$; structure for near-complete or dense graphs is less well-understood.
- Extension to spectral extremal problems for the second largest eigenvalue, spectral gaps, and non-adjacency matrices (e.g., normalized Laplacian) is largely undeveloped [2511.06643].
- For minimization in fixed-invariant classes (such as matching number), rich combinatorial phenomena linking quasi-adjacency, domination, and balancing constraints suggest broader block-graph structural classifications [2601.18223].

## 7. Position and Impact in Extremal Graph Theory

The Brualdi–Solheid Problem has unified classical combinatorial extremal theory with algebraic (spectral) methods, revealing strong connections between global spectral parameters and fine graph structure. For a wide array of forbidden subgraphs and invariants, extremal graphs in the spectral setting are determined and often coincide with their combinatorial (edge-count extremal) counterparts, although spectral analysis sometimes uncovers exceptional extremals inaccessible to purely combinatorial arguments [2305.03952][2302.03313][2007.06922][2511.06643][2412.16019][2405.20056][2601.18223]. This suggests an underlying universality and hints at further algebraic symmetries in graph structure yet to be fully described.

Source: https://www.emergentmind.com/topics/brualdi-solheid-problem