ℓ₁-Fiedler Value

• ℓ₁-Fiedler value is a graph invariant defined as the minimum ℓ₁-smoothness over zero-sum, unit ℓ₁-norm vectors, directly linking spectral connectivity with combinatorial expansion.
• It offers a combinatorial analogue to algebraic connectivity by reformulating the sparsest-cut problem with a variational principle that optimizes edge differences.
• Sharp inequalities connect b(G) with Laplacian eigenvalues and isoperimetric parameters, underpinning applications in extremal tree constructions and NP-hard optimization.

The $\ell_1$-Fiedler value $b(G)$ is a combinatorial graph invariant introduced as an $\ell_1$-norm analogue of graph algebraic connectivity. Formally, $b(G)$ is defined for a simple undirected graph $G=(V,E)$ as the minimum $\ell_1$-smoothness over all zero-sum, unit $\ell_1$-norm vectors, optimizing the sum of edge differences. This parameter provides a direct connection to the sparsest-cut problem and exposes new relationships between spectral graph theory and combinatorial expansion properties, with deep ties to Laplacian eigenvalues, isoperimetric numbers, and extremal graph constructions (Andrade et al., 2023, Kannan et al., 9 Jan 2026).

1. Definition and Variational Formulation

Let $G=(V,E)$, $|V|=n$, and $x\in\mathbb{R}^n$ with $b(G)$0, $b(G)$1. The $b(G)$2-Fiedler value is: $b(G)$3 Any optimal $b(G)$4 is called an $b(G)$5-Fiedler vector. This minimization can alternatively be formulated using nonnegative vectors $b(G)$6 with orthogonal supports and equal sums: $b(G)$7 This variational principle establishes $b(G)$8 as a combinatorial analogue to the second Laplacian eigenvalue $b(G)$9, substituting the quadratic $\ell_1$0 smoothing with $\ell_1$1 total variation (Andrade et al., 2023, Kannan et al., 9 Jan 2026).

2. Connectivity and Sparsest-Cut Equivalence

A central property (Theorem 2) is that $\ell_1$2 if and only if $\ell_1$3 is connected. If $\ell_1$4 is disconnected, a feasible $\ell_1$5 supported on a component yields zero edge contributions, so $\ell_1$6. When $\ell_1$7 is connected, any nonzero $\ell_1$8 forces at least one edge to have nonzero difference (Andrade et al., 2023).

Crucially, $\ell_1$9 has an explicit combinatorial characterization as the edge-density of the sparsest cut: $b(G)$0 If a set $b(G)$1 achieves the minimum, the corresponding $b(G)$2-Fiedler vector is: $b(G)$3 This equivalence directly connects $b(G)$4 to the classic sparsest-cut problem and leverages cut-based expansion tools for analysis (Andrade et al., 2023, Kannan et al., 9 Jan 2026).

3. Fundamental Inequalities and Bounds

Multiple sharp inequalities relate $b(G)$5 to spectral, degree, and isoperimetric invariants:

• $b(G)$6, where $b(G)$7 is the algebraic connectivity and $b(G)$8 the largest Laplacian eigenvalue.
• $b(G)$9 for $G=(V,E)$0.
• Vertex-degree bound: $G=(V,E)$1, with $G=(V,E)$2 the minimum degree.
• Isoperimetric lower bound: $G=(V,E)$3, $G=(V,E)$4 (Cheeger constant).

Nordhaus-Gaddum inequalities (Theorem 3.7 (Kannan et al., 9 Jan 2026)) constrain the sum of $G=(V,E)$5 and $G=(V,E)$6: $G=(V,E)$7 with equality characterized by complete graphs and limiting cases for stars and their complements (Kannan et al., 9 Jan 2026).

4. Explicit Formulas and Extremal Constructions for Trees

For trees $G=(V,E)$8, $G=(V,E)$9 admits a closed formula: $\ell_1$0 where $\ell_1$1 is a center-edge minimizing the partition size differential. Key special cases:

• Paths: $\ell_1$2 if $\ell_1$3 is even; $\ell_1$4 if $\ell_1$5 is odd.
• Stars: $\ell_1$6.

Extremal tree results (Theorem 4.10 (Kannan et al., 9 Jan 2026)) show the star graph $\ell_1$7 globally maximizes $\ell_1$8 among $\ell_1$9-vertex trees; the path $\ell_1$0 minimizes it. Prescribed tree parameters (diameter $\ell_1$1, maximum degree $\ell_1$2, number of pendant vertices $\ell_1$3) yield construction schemes and explicit bounds for $\ell_1$4 (Kannan et al., 9 Jan 2026).

5. Laplacian and Edge-Connectivity Connections

Given $\ell_1$5 as the sparsest-cut subset and the induced $\ell_1$6-Fiedler vector $\ell_1$7, summations of Laplacian products yield

$\ell_1$8

If the edge-connectivity of $\ell_1$9 is $G=(V,E)$0, then

$G=(V,E)$1

Equality holds if and only if $G=(V,E)$2 may be constructed by adding $G=(V,E)$3 edges incident solely to an isolated vertex such that at each intermediate step, the singleton induces the unique sparsest cut (Kannan et al., 9 Jan 2026).

6. Addition of Pendant Vertices

Successively attaching $G=(V,E)$4 pendant vertices to $G=(V,E)$5 yields

$G=(V,E)$6

Each pendant attachment reduces $G=(V,E)$7 multiplicatively by at most $G=(V,E)$8 ($G=(V,E)$9 the number of vertices at step $|V|=n$0) (Kannan et al., 9 Jan 2026).

7. Relationship with Isoperimetric Number and Cheeger-Type Bounds

The isoperimetric number $|V|=n$1 upper bounds $|V|=n$2: $|V|=n$3 Combining with known bounds for $|V|=n$4, such as Mohar's $|V|=n$5, one obtains

$|V|=n$6

For $|V|=n$7-regular graphs, Cheeger’s inequality in the $|V|=n$8 setting gives $|V|=n$9. When the minimum sparsest cut is of size $x\in\mathbb{R}^n$0 and $x\in\mathbb{R}^n$1 is even, $x\in\mathbb{R}^n$2; singleton minimizers yield $x\in\mathbb{R}^n$3 (Kannan et al., 9 Jan 2026).

8. Computational Complexity and Norm Variants

Computing $x\in\mathbb{R}^n$4 and associated $x\in\mathbb{R}^n$5-Fiedler vectors is NP-hard, as it is equivalent to the sparsest cut problem. In contrast, the $x\in\mathbb{R}^n$6-Fiedler value

$x\in\mathbb{R}^n$7

admits a polynomial-time solution through $x\in\mathbb{R}^n$8 linear programs. For paths, $x\in\mathbb{R}^n$9 (Andrade et al., 2023).

---

The $b(G)$00-Fiedler value $b(G)$01 thus bridges spectral and combinatorial connectivity, underpinning sparsest cut duality, tree extremal constructions, edge-connectivity, and isoperimetric bounds in both classical and parameterized graph families. Its direct link to NP-hard optimization and Cheeger-type inequalities positions $b(G)$02 as a central object in modern combinatorial spectral theory (Andrade et al., 2023, Kannan et al., 9 Jan 2026).