Refined Modulus Bounds

• Refined modulus bounds are precise inequalities linking modulus quantities to intrinsic invariants such as geometry, structure, and analytical characteristics.
• They sharpen classical results by introducing sharper constants, explicit parameter dependencies, and new structural features like directionality and duality.
• Applications span geometric function theory, higher-dimensional mappings, polynomial estimates, operator theory, and adaptive algorithms in signal processing.

A refined modulus bound is any precise, often optimal, inequality or estimate relating a modulus quantity—be it the modulus of a family of curves, surfaces, continuity, algebraic/logarithmic roots, or functionals—to more intrinsic invariants, geometry, or structure. These bounds sharpen classical results by providing sharper constants, explicit dependence on parameters, or the inclusion of new structural features (directionality, duality, dynamic adaptation, etc.) By capturing finer geometric or analytic details, they serve as tools for analyzing extremality, stability, regularity, or worst-case performance in diverse areas across analysis, geometry, combinatorics, probability, and mathematical physics.

1. Duality and Sharp Modulus Reciprocity

The duality principle for modulus—originally from geometric function theory—relates the $p$-modulus of a family of curves ($p$-energy minimization) to a dual $q$-modulus over “surface” objects, with $1/p + 1/q = 1$. The sharpest lower bound to date is given by the following inequality in a compact metric space $(X, d)$ of finite Hausdorff $N$-measure, for any pair of disjoint continua $E, F \subset X$:

$$\bigl(_p\Gamma(E, F)\bigr)^{1/p} \bigl(_q\Sigma_H(E, F)\bigr)^{1/q} \geq \frac{v_N}{2v_{N-1}}$$

where $v_k = \pi^{k/2}/\Gamma(k/2 + 1)$, $\Gamma(E, F)$ denotes the family of rectifiable paths connecting $E$ and $F$, and $\Sigma_H$ is the corresponding family of separating hypersurfaces. This result is best possible and realized for the rectangle $([0, 1]^2, \|\cdot\|_\infty)$, answering a question of Rajala and Romney. The proof involves a Lipschitz approximation and a refined co-area argument, leading to tightness for planar quadrilaterals: the minimal possible constant is $v_2/(2v_1) = \pi/4$ (Eriksson-Bique et al., 2021).

2. Refined Modulus Bounds in Higher-Dimensional Mappings and Cavitation

When studying homeomorphisms in $\mathbb{R}^n$, especially phenomena such as cavitation—mapping a punctured ball to a nondegenerate ring—classical modulus bounds using outer ($K_f$) and inner ($L_f$) dilatation often fail to distinguish cavitation from continuous extension. Two refined directional dilatations are thus introduced:

• Angular dilatation: $D_f(x, x_0) = J_f(x) / \ell_f(x, x_0)^n$ measures the distortion orthogonal to the radial direction.
• Normal (directional) dilatation: $$Q_f(x, x_0) = (|\partial_u f(x)|^n / J_f(x))^{1/(n-1)}$$ quantifies minimal radial stretching relative to the Jacobian.

Optimal modulus estimates for the image of the standard condenser $A(r, R) = \{x: r < |x| < R\}$ are:

$$\int_{S^{n-1}} [ I_Q(u; r, R) ]^{1-n} d\sigma(u) \leq M(f(\Gamma)) \leq \left[ \int_r^R \Phi_D(t)^{1/(1-n)} dt \right]^{1-n}$$

where $I_Q(u; r, R) = \int_r^R Q_f(tu) dt/t$ and $$\Phi_D(t) = \int_{S^{n-1}} D_f(tu) t^{n-1} d\sigma(u)$$. These bounds are strictly stronger than the classical ones and deliver both necessary and sufficient integral criteria for the extension vs. cavitation dichotomy, as demonstrated by explicit examples (Golberg et al., 21 Dec 2025).

3. Explicit Lower Bounds for Path-Family $p$-Moduli and Boundary Finiteness

For the family $\Gamma(A, A^*, D)$ of all locally rectifiable paths in a domain $D \subset \mathbb{R}^n$ connecting two non-degenerate continua, the $p$-modulus $M_p(\Gamma)$ with $p > n-1$ obeys the explicit bound:

$$M_p(\Gamma(A, A^*, D)) \geq \frac{n-p}{2^n K_{n,p} b_{n,p}} \left((2r)^{n-p} - r^{n-p}\right)$$

where $r$ is determined by the minimal separation and distance to the boundary, and constants $K_{n,p}, b_{n,p}$ depend only on $n, p$. Furthermore, if $D$ has $p$-strongly accessible boundary, then there exists a uniform $\varepsilon^* > 0$ such that all path-families intersecting concentric neighborhoods near the boundary have $p$-modulus at least $\varepsilon^*$, and $D$ is finitely connected at the boundary (Sevost'yanov et al., 2024).

4. Refined Modulus Estimates for Polynomial Roots

For the monic polynomial $p(z)$ of degree $n$, root-modulus bounds classical in numerical analysis (Cauchy, Fujiwara) have worst-case overestimation factors proportional to $n$ but with large constants. The nearly optimal absolute bound,

$R(p) \leq I(p)$

with $I(p)$ a function (explicitly constructed from four Cassini-oval families of the companion matrix) that satisfies

$\sup_{\deg n} I(p) / R(p) \leq 1.4655\, n,$

attains the van der Sluis theoretical optimum $1/(2^{1/n}-1) \sim 1.442\, n$ up to at most $2\%$ as $n \to \infty$. This is the best known explicit bound with respect to worst-case overestimation for all polynomials of large degree and is constructed via the companion-matrix similarity approach (Batra, 2024).

Bound	Asymptotic Growth Rate	Worst-case Factor
van der Sluis	$1/(2^{1/n}-1)$	$\sim1.442\, n$
Fujiwara	$2\, n$	$2\, n$
Improved Lagrange	$1.58\, n$	$1.58\, n$
New bound $I(p)$	$1.4655\, n$	$1.4655\, n$

5. Refined Modulus Bounds in Operator Theory, Harmonic Analysis, and PDE

In the regularity theory of maximal operators and functional inequalities:

• Hardy-Littlewood maximal function: The best constant for the modulus of continuity of the uncentered maximal operator on $\operatorname{Lip}_\alpha$ for $0<\alpha\leq 1$ is $1/(1+\alpha)$ on intervals, and $(\sqrt{2}-1)$ in $L^1$ on $\mathbb{R}$; in higher dimensions, for Euclidean balls, the sharp bound is $2^{-\alpha/q}$ with $q$ the dual exponent (Aldaz et al., 2010).
• Spectral gaps of combinatorial Laplacians: For any strongly convex subgraph $S$ of an invariant homogeneous graph of diameter $D$, the first nonzero eigenvalue (the gap) of the Laplacian satisfies

$$\lambda_1(L_S) \geq 2\left(1-\cos\frac{\pi}{D+1}\right)$$

achieving equality for path graphs, and with further refinements in the presence of log-concavity or additional convexity (Jarret et al., 2015).

6. Refined Modulus in Complex Dynamics and Model-Independent Function Bounds

• Polynomial-like maps (Pommerenke-Levin-Yoccoz inequality): If $P:U_1\to U_0$ is a polynomial-like map of degree $d$, and the modulus satisfies $\mu = \mathrm{mod}(U_0 \setminus U_1)>0$, then there is a sharp upper bound, e.g.

$\frac{1}{\mu} \geq \frac{\pi |Z|}{\log D}$

where $|Z|$ is the number of legal cuts, with equality in the absence of bounded Fatou accesses (Blokh et al., 2022).

• Analytic function theory (model-independent bounds): For the modulus of the pion form factor $F_\pi(t)$, maximum modulus principles combined with $L^2$ norm constraints, the Fermi–Watson phase, and high-energy data yield optimal, model-independent upper and lower bounds via the solution to a Meiman interpolation problem for functions analytic in the unit disk, with explicit determinantal constraints and explicit outer functions (Ananthanarayan et al., 2012).

7. Dynamic and Adaptive Modulus Bounds in Stochastic Algorithms

• Adaptive bounds in signal processing: The set-membership constant modulus algorithm with generalized sidelobe canceler (SM-CM-GSC) adapts a dynamic modulus bound $\gamma(i)$ based on both parameter-dependent and interference-dependent terms: $$\gamma(i+1) = (1-\rho)\gamma(i) + \rho [\sqrt{\psi\nu(i)} + \sqrt{\lambda \|\hat{w}(i)\|^2 \hat{\sigma}_n^2(i)}]$$ This dynamic bound improves convergence rates and steady-state performance, sharply reducing the update rate in adaptive beamforming scenarios, and is provably optimal in tracking sudden changes in environment compared to fixed or static modulus bounds (Cai et al., 2014).

---

Refined modulus bounds thus formalize and optimize the intrinsic relationship between geometric, analytic, or spectral invariants and the structural properties of objects or operators of interest, using precise, context-sensitive estimates that strictly improve upon classical results. Their applications are pervasive: from moduli in metric geometry and function theory, through harmonic analysis and operator theory, to algorithmic adaptation in engineering and number-theoretic combinatorics. Their unifying feature is the deployment of sharper constants, attention to symmetry and directionality, and adaptation to additional structure or environmental information, enabling both sharper theoretical insight and more effective practical performance.