On Computation of Capacities and Conformal Invariants

Basic Information

• Paper ID: 2507.11648
• Title: On computation of capacities and conformal invariants
• Authors: Mohamed M. S. Nasser, Matti Vuorinen
• Classification: math.CV (Complex Variables)
• Publication Date: November 20, 2025 (arXiv v2)
• Paper Link: https://arxiv.org/abs/2507.11648

Abstract

This paper surveys the computation of conformal capacity, generalized capacity, and logarithmic capacity of planar condensers, emphasizing recent work by the authors during 2020-2025. The article discusses boundary integral equation methods based on generalized Neumann kernels for computing various conformal invariants, including harmonic measure, quadrilateral modulus, reduced modulus, hyperbolic capacity, and elliptic capacity. The solution of mixed Dirichlet-Neumann boundary value problems for Laplace's equation plays a key role in the methodology. The paper concludes with a thematic bibliography of constructive complex analysis and potential theory.

Research Background and Motivation

Research Problem

Conformal capacity is one of the core concepts in potential theory with widespread applications in geometric function theory, partial differential equations, and the study of conformal invariants. However, exact values of condenser capacity are known only in very few cases, necessitating effective numerical computation methods.

Problem Significance

1. Theoretical Importance: Conformal capacity is intimately related to the modulus of curve families and is a fundamental tool in geometric function theory
2. Historical Continuity: Inherits pioneering work from mathematicians including Koebe, Bergman, Grötzsch, Teichmüller, and Ahlfors
3. Computational Necessity: The classical work of Pólya and Szegö posed numerous isoperimetric problems requiring numerical solution methods
4. Applied Value: Has practical applications in physics, engineering, computer graphics, and fluid dynamics

Limitations of Existing Methods

1. Scarcity of Analytical Formulas: Exact formulas exist only for a few special geometric shapes
2. Complexity of Traditional Methods: Classical methods such as Schwarz-Christoffel mappings are impractical for manual computation
3. Computational Efficiency: Early numerical methods were slow and had limited accuracy
4. Limited Applicability: Different methods apply to different types of domains, lacking a unified framework

Motivation for This Work

The authors aim to provide a unified numerical framework based on boundary integral equation methods that can efficiently and accurately compute various conformal invariants, systematically summarizing research results from the past five years.

Core Contributions

1. Unified Computational Framework: Proposes a boundary integral equation method based on generalized Neumann kernels applicable to a broad range of conformal invariant computations
2. Efficient Algorithm Implementation: Combines fast multipole method (FMM) and generalized minimal residual method (GMRES), achieving O((m+1)n log n) computational complexity
3. Multi-Invariant Computation: Systematically demonstrates method applications to:
   • Generalized condenser capacity
   • Logarithmic capacity
   • Hyperbolic and elliptic capacity
   • Reduced modulus
   • Quadrilateral modulus
   • Harmonic measure
   • Hyperbolic distance
4. Complex Domain Handling: Method handles domains with corner points, nearly touching boundaries, non-convex boundaries, and high connectivity
5. Systematic Literature Review: Provides thematic bibliography for constructive complex analysis and potential theory
6. Numerical Verification: Systematic comparison with hp-FEM and other methods, achieving accuracy at 10^(-13) level

Methodology Details

Problem Definition

Given an (m+1)-connected region G (bounded by m+1 smooth Jordan curves Γ_k, k=0,1,...,m), compute various conformal invariants such as capacity, modulus, and harmonic measure.

Input: Parametric representation of region boundary η_k(t) Output: Numerical approximation of desired conformal invariants Constraints: Boundary curves are smooth or piecewise smooth

Core Method Architecture

1. Boundary Integral Equation Foundation

Generalized Neumann Kernel is defined as:

N(s,t) = (1/π) Im[A(s)/A(t) · η'(t)/(η(t)-η(s))]

where A(t) is a complex function:

• Bounded domain: A(t) = η(t) - α (α is an auxiliary point in the domain)
• Unbounded domain: A(t) = 1

Core Integral Equation (Theorem 3.1):

(I - N)ρ = -Mγ

where:

• ρ is the unknown function
• M is a singular integral operator
• γ is a given Hölder continuous function
• N is a compact operator

The solution ρ and piecewise constant function ν determine the values of analytic function f on the boundary:

f(η(t)) = [γ(t) + ν(t) + iρ(t)]/A(t)

2. Numerical Solution Method

Discretization: Uses Nyström method with trapezoidal rule

• For C^∞ boundaries: convergence rate O(e^(-cn))
• For C^(q+2) boundaries: convergence rate O(1/n^q)

Linear System Solution:

• GMRES method solves discretized (m+1)n×(m+1)n linear system
• FMM accelerates matrix-vector products
• Tolerance set to 10^(-14)

Corner Point Handling: Uses graded mesh method

• Boundary parametrization: η(t) = η̂(δ(t))
• δ(t) function eliminates discontinuities in solution derivatives at corners
• Grading parameter p=3, convergence rate O(n^(-3))

Specific Applications

1. Generalized Condenser Capacity Computation

For condenser C = (Ω, E, δ), capacity is defined as:

cap(C) = inf_{u∈A} ∫_Ω |∇u|² dm

Computational Steps:

1. For each inner boundary Γ_k, solve integral equation to obtain μ_k and ν_k
2. Solve (m+1)×(m+1) linear system to determine constants a_1,...,a_m, c
3. Capacity calculation: cap(C) = 2π Σ_^m δ_k a_k

2. Logarithmic Capacity Computation

For compact set E, logarithmic capacity is defined as:

cap_l(E) = lim_{z→∞} exp(log|z| - g_G(z))

where g_G is the Green function.

Computational Method:

1. Construct lemniscatic domain Ω = {z: |U(z)| > κ}
2. Solve (m+1)×(m+1) linear system to determine ℓ_0,...,ℓ_m and log(κ)
3. Logarithmic capacity equals κ

3. Quadrilateral Modulus Computation

For quadrilateral (G; z_1, z_2, z_3, z_4), modulus h satisfies the existence of conformal mapping to rectangle 0,1×0,h.

Two-Step Method:

1. Compute conformal mapping Φ_1: G → B² (unit disk)
2. Use cross-ratio formula: k = |ŵ_1, ŵ_2, ŵ_3, ŵ_4|
3. Modulus: h = (2/π)μ(1/√k)

Technical Innovations

1. Unified Framework: Single integral equation method applies to diverse conformal invariants with minimal code modification
2. Efficient Implementation: FMM+GMRES combination significantly improves computational speed
3. Flexibility: Easy to modify code for different domain types
4. Accuracy Guarantee: Precision comparable to or exceeding hp-FEM methods
5. Corner Handling: Innovative graded mesh method for non-smooth boundaries

Experimental Setup

Test Domain Types

1. Annular domains: Regions between concentric circles
2. Nested squares: Concentric square regions
3. Disk with polygonal holes: Unit disk with regular polygon removed
4. Lens-shaped plates: Domains bounded by two circular arcs
5. Multiply-punctured disks: Disks with multiple small circles removed
6. Elliptical interiors/exteriors: Elliptical domain regions
7. Trapezoids and gear domains: Polygonal regions
8. Amoeba-shaped domains: Complex parametrized boundaries

Evaluation Metrics

1. Relative Error: |computed value - exact value|/|exact value|
2. Convergence Rate: Error variation with mesh point number n
3. Computation Time: Algorithm execution time
4. Comparison with Other Methods: Precision comparison with hp-FEM, AFEM
5. Verification Tests: Such as mod(G; z_1,z_2,z_3,z_4)·mod(G; z_4,z_1,z_2,z_3) ≈ 1

Implementation Details

• Programming Language: MATLAB
• Mesh Points: Typically n = 2^8 to 2^14
• GMRES Tolerance: 10^(-14)
• FMM Tolerance: 0.5×10^(-15)
• Maximum Iterations: 100
• Grading Parameter: p = 3 (corner cases)
• Code Repository: https://github.com/mmsnasser/cap

Experimental Results

Main Results

1. Annular Capacity (Example 4.1)

• Domain: Unit disk with disk of radius a removed
• Exact Formula: cap = 2π/log(1/a)
• Convergence: Exponential convergence (smooth boundary)
• Accuracy: Relative error reaches machine precision

2. Nested Squares (Example 4.2)

• Domain: (-1,1)×(-1,1) with -a,a×-a,a removed
• Exact Formula: cap = 4π/μ(r) (r determined by complex formula)
• Convergence: Algebraic convergence O(n^(-3)) (corners)
• Verification: Consistent with literature results 21, 132

3. Lens-Shaped Plate (Example 4.4)

• Domain: Unit disk with lens-shaped domain removed
• Special Cases: s=0 gives cap=2π/μ(2a/(1+a²)), s=a gives cap=2π/log(1/a)
• Estimate Formula: cap ≈ 2π/log(2(π-θ)/(πa))
• Results: Estimate values align well with computed values when s approaches a

4. Elliptical Interior Reduced Modulus (Example 7.1)

• Domain: Interior of ellipse η(t) = cosh(r+it)
• Exact Formula: m(G,0) = (1/2π)log(π/(2√sK(s))), s=μ^(-1)(2r)
• Relative Error: Reaches 10^(-14) level at n=2^8
• Contour Lines: Maximum appears at α=0

5. Trapezoid Modulus (Example 8.1)

• Domain: Trapezoid with vertices 0,1,1+iL,i(L-1)
• Exact Formula: mod = π/(2μ(k)) (k determined by complex formula)
• L=1.5, n=2^12: Relative error 5.47×10^(-14)
• Verification: Consistent with literature 56, 132

6. Hyperbolic/Elliptic Capacity (Examples 6.1, 6.2)

• Domain: Unit disk with ellipse removed
• Symmetry Verification: When E=-E, cap_h(E)=cap_e(E)
• Numerical Verification: |cap_h - cap_e| = 9.99×10^(-16)
• n=2^10: Computed value 0.634497711721981/982

Comparison with Other Methods

hp-FEM Comparison (Literature 56, 57, 58)

• Consistency: Numerical results consistent at 10^(-13) level
• Speed: Boundary integral method typically faster
• Flexibility: Boundary integral method easier to modify

AFEM Comparison (Literature 21, 143)

• Accuracy: Consistent with Samuelsson's results at 10^(-6) level
• Grötzsch Capacity: γ_3(r) computation results align

SCToolbox Comparison

• Simply Connected Polygons: Precision close to SCToolbox
• Multiply Connected Domains: Boundary integral method shows advantages

Ablation Studies

While the paper lacks an explicit ablation study section, different domain types demonstrate various method aspects:

1. Smooth vs. Corners: Circular domains (exponential convergence) vs. polygonal domains (algebraic convergence)
2. Connectivity: Computation from simply connected to 5-connected domains
3. Mesh Density: Convergence behavior at different n values
4. FMM Acceleration: Computational complexity reduction from O(n²) to O(n log n)

Experimental Findings

1. Universality: Single method applies to various conformal invariants with minimal code changes
2. Robustness: Handles nearly touching boundaries, non-convex boundaries, and high-connectivity regions
3. Accuracy: Precision comparable to or exceeding other mature methods
4. Efficiency: FMM acceleration significantly improves computational speed
5. Visualization: Abundant contour plots and convergence curves provided

Related Work

Classical Theoretical Foundations

1. Conformal Mapping Theory: Koebe, Riemann, Schwarz-Christoffel
2. Capacity Theory: Pólya-Szegö 133, Ahlfors-Beurling 6
3. Modulus Theory: Grötzsch, Teichmüller, Fuglede 41
4. Potential Theory: Landkof 93, Ransford 138, Tsuji 159

Numerical Method Development

Early Methods (1950-1980)

• Gaier 42: Pioneering work in constructive methods
• Henrici 69: Systematic numerical conformal mapping theory
• Pólya-Szegö 133: Numerical tables for isoperimetric problems

Modern Methods (1980-2000)

• Schwarz-Christoffel Method: Trefethen 152, Driscoll-Trefethen 37
• Kerzman-Stein Method: Integral equations based on Szegö kernel
• FMM Development: Greengard-Rokhlin 53

Recent Progress (2000-2025)

• FEM Methods: Samuelsson 143, Hakula-Rasila-Vuorinen 60-62
• Boundary Integral Methods: Wegmann 167, Nasser 109-115
• Circle Domain Methods: DeLillo et al. 34, Nasser 113
• Zipper Method: Marshall-Rohde 103

Relationship to Related Work

1. Inherits Wegmann's Work: Develops generalized Neumann kernel method
2. Complements hp-FEM: Systematic precision comparison and verification
3. Extends Application Scope: From conformal mappings to multiple invariants
4. Provides Unified Framework: Integrates multiple computational tasks

Advantages of This Work

1. Method Unification: Single framework solving multiple problems
2. Computational Efficiency: FMM acceleration significantly improves speed
3. Broad Applicability: Handles various complex domains
4. Precision Verification: Systematic comparison with other methods
5. Literature Review: Complete thematic bibliography

Conclusions and Discussion

Main Conclusions

1. Method Effectiveness: Boundary integral equation method efficiently and accurately computes various conformal invariants
2. Unified Framework: Single method applies to capacity, modulus, harmonic measure, and other problems
3. Computational Precision: Precision comparable to hp-FEM methods, reaching 10^(-13) in some cases
4. Computational Speed: FMM acceleration typically provides speed advantage over other methods
5. Flexibility: Easy to modify for different domains and invariants

Limitations

1. Theoretical Completeness: Theoretical analysis insufficient for certain extreme cases (e.g., extremely narrow channels)
2. Parameter Selection: Auxiliary point α selection affects certain problems; lacks automatic selection strategy
3. Three-Dimensional Extension: Method primarily addresses planar problems; 3D extension requires further research
4. Analytical Solution Verification: Many examples lack exact solutions for verification
5. Error Estimates: Mostly experimental error estimates; lacks theoretical error bounds

Future Directions

Research directions explicitly proposed in the paper:

1. Fundamental Frequency of Polygonal Domains: Estimate fundamental frequency of Laplace operator in bounded polygonal regions
2. Grötzsch-Teichmüller Problem: Study λ_G and μ_G in polygonal planar regions
3. Three-Dimensional Capacity: Extend method to three-dimensional Grötzsch capacity γ_3(r)
4. Isoperimetric Problems: Inherit Pólya-Szegö's work, study extremal problems for domain functionals
5. Adaptive Methods: Develop adaptive mesh and parameter selection strategies

In-Depth Evaluation

Strengths

1. Method Innovation

• Unified Framework: First systematic demonstration of single method handling such diverse conformal invariants
• Technical Integration: Skillfully combines integral equations, FMM, and GMRES
• Corner Handling: Innovative graded mesh method for non-smooth boundaries

2. Experimental Sufficiency

• Rich Case Studies: Covers circular, polygonal, elliptical, and other domain types
• Systematic Comparison: Compares with hp-FEM, AFEM, and other methods
• Complete Visualization: Abundant charts showing convergence and contour lines
• Open Code: GitHub repository provided for reproducibility

3. Result Convincingness

• Precision Verification: Multiple verification approaches (exact solutions, other methods, theoretical relationships)
• Convergence Analysis: Clear demonstration of exponential convergence for smooth boundaries, algebraic for corners
• Consistency Checks: Uses theoretical relationships (e.g., modulus reciprocal relations) for verification

4. Writing Quality

• Clear Structure: 12 chapters with logical progression from theory to applications
• Literature Review: Thematic bibliography (sections 12.1-12.12) highly valuable as reference
• Mathematical Rigor: Theorems, definitions, and formulas precisely stated
• Readability: Abundant examples and figures aid understanding

Weaknesses

1. Method Limitations

• Two-Dimensional Restriction: Method primarily addresses planar problems; 3D extension unclear
• Smoothness Requirements: While handling corners, still requires boundary regularity
• Parameter Dependence: Auxiliary point α selection impact insufficiently discussed; lacks automatic selection strategy
• Convergence Theory: Convergence analysis for non-smooth boundaries not sufficiently deep

2. Experimental Setup

• Error Estimates: Mostly experimental errors; lacks theoretical error bounds
• Extreme Cases: Insufficient testing of extreme geometries (extremely narrow channels, near-degenerate cases)
• Computation Time: Actual computation time comparison between methods not systematically reported
• Large-Scale Testing: Insufficient systematic testing for high connectivity (m>10)

3. Theoretical Analysis

• Optimality: Method optimality (convergence rate, computational complexity) not proven
• Stability: Numerical stability theoretical analysis insufficient
• Parameter Selection: Lacks theoretical guidance for auxiliary point, mesh density selection
• Failure Conditions: Boundary conditions for method failure not clearly defined

4. Application Depth

• Real-World Problems: Lacks actual application cases from physics and engineering
• Software Tools: While code is public, lacks user-friendly software package
• User Guide: Lacks detailed usage guide for non-specialist users
• Parallelization: Parallel computing and GPU acceleration not discussed

Impact

1. Field Contribution

• Methodology: Provides unified efficient framework for conformal invariant computation
• Theory: Connects integral equation theory with conformal mapping computation
• Applications: Provides powerful numerical tool for geometric function theory
• Literature: Thematic bibliography will become important reference resource

2. Practical Value

• Computational Efficiency: Fast and accurate method suitable for practical computation needs
• Flexibility: Easy to modify for new problems
• Extensibility: Framework extensible to other conformal invariants
• Educational Value: Rich examples suitable for teaching and learning

3. Reproducibility

• Open Code: Complete implementation in GitHub repository
• Clear Description: Algorithm steps and parameter settings clearly described
• MATLAB Platform: Uses widely-available MATLAB
• Test Cases: Abundant examples for implementation verification

4. Expected Impact

• Citation Value: Survey nature will make it important field reference
• Method Promotion: Method likely applied to more fields
• Software Development: May catalyze development of more complete software tools
• Follow-up Research: Provides foundation for 3D extension, theoretical analysis

Applicable Scenarios

1. Ideal Applications

• Conformal Invariant Computation: Research requiring various capacity and modulus calculations
• Geometric Function Theory: Theoretical work studying conformal mapping properties
• Isoperimetric Problems: Domain functional optimization and extremal problems
• Potential Theory: Harmonic function and Green function computation

2. Domain Types

• Smooth Boundaries: Exponential convergence, extremely high precision
• Polygonal Domains: Algebraic convergence, good precision
• Multiply Connected: Method naturally suited for multiply connected domains
• Complex Geometry: Non-convex, nearly touching, and other complex cases

3. Computational Needs

• High Precision: Computations requiring 10^(-10) or higher precision
• Batch Computation: Computing invariants for numerous different domains
• Parameter Study: Studying invariant variation with geometric parameters
• Verification Tool: Verifying results from other methods (e.g., FEM)

4. Unsuitable Scenarios

• Three-Dimensional Problems: Method not extended to 3D
• Extreme Geometry: Extremely narrow channels, degenerate cases may fail
• Real-Time Computation: While fast, may not meet real-time requirements
• Low Precision Requirements: Simple methods may be more appropriate when high precision unnecessary

Selected Important References

Classical Foundations

1. Ahlfors-Beurling (1950): Foundational work on modulus theory
2. Pólya-Szegö (1951): Classical treatise on isoperimetric problems
3. Fuglede (1957): Extension of extremal length theory to higher dimensions

Numerical Methods

4. Gaier (1964): Constructive conformal mapping methods
5. Henrici (1986): Computational Complex Analysis, Volume 3
6. Driscoll-Trefethen (2002): Schwarz-Christoffel mappings

Potential Theory

7. Ransford (1995): Potential theory in the complex plane
8. Garnett-Marshall (2008): Harmonic measure
9. Hariri-Klén-Vuorinen (2020): Conformal invariant metrics

Authors' Prior Work

10. Nasser (2009, 2015): Development of boundary integral equation methods
11. Hakula-Nasser-Vuorinen (2023): hp-FEM and integral equation comparison
12. Nasser-Vuorinen (2020-2021): Capacity and conformal invariant computation

Overall Assessment: This is a high-quality survey research paper systematically demonstrating the power of boundary integral equation methods for conformal invariant computation. The method is unified, experiments are comprehensive, writing is clear, and the thematic bibliography is highly valuable. Main contributions lie in providing an efficient, accurate, and flexible unified computational framework. Limitations include two-dimensional restriction and theoretical analysis depth. Will have significant impact on geometric function theory and computational mathematics, particularly providing powerful tools for numerical conformal mapping.