Simple Gradient Flow Equation for the Bounce Solution

Basic Information

• Paper ID: 1907.02417
• Title: Simple Gradient Flow Equation for the Bounce Solution
• Author: Ryosuke Sato (DESY)
• Classification: hep-ph astro-ph.CO hep-th math-ph math.MP
• Publication Date: July 2019 (Final version October 2025)
• Paper Link: https://arxiv.org/abs/1907.02417

Abstract

Inspired by recent work of Chigusa, Moroi, and Shoji, this paper proposes a novel simple gradient flow equation for deriving the bounce solution that contributes to false vacuum decay. The research is based on the theoretical framework of Coleman, Glaser, and Martin, solving the problem through minimizing kinetic energy under the constraint of fixed potential energy. The bounce solution is obtained through a scale transformation of the solution to this problem. The convergence of this method is demonstrated to be robust with respect to the choice of initial configuration.

Research Background and Motivation

Core Problem

False vacuum decay is an important problem in particle physics and cosmology. The decay rate of false vacuum can be calculated from the "imaginary part" of the Euclidean path integral, where the dominant contribution comes from the non-trivial solution with minimal action—the bounce solution.

Problem Significance

1. Physical Importance: The bounce solution plays a crucial role in the false vacuum decay process, directly affecting the calculation of decay rates
2. Technical Challenge: The bounce solution is a saddle point of the action rather than a stable minimum, making direct calculation difficult
3. Multi-field Complexity: Calculations become particularly complex for multi-scalar field models

Limitations of Existing Methods

Existing algorithms include modified gradient flow, modified action making the bounce solution a local minimum, gradually changing friction coefficients, machine learning methods, etc., but all suffer from convergence or computational efficiency issues.

Research Motivation

Based on recent work by Chigusa et al., this paper aims to propose a simpler and more stable gradient flow equation utilizing the Coleman-Glaser-Martin (CGM) theoretical framework, transforming the problem into a constrained optimization problem.

Core Contributions

1. Proposes New Gradient Flow Equation: A novel flow equation based on CGM simplification, avoiding direct treatment of saddle points
2. Theoretical Guarantees: Proves convergence and robustness to initial conditions
3. Efficient Implementation: The developed numerical package computes Euclidean bounce action for 1-8 scalar field models with O(0.1)% accuracy in O(0.1) seconds
4. Extensive Verification: Validates the method's effectiveness through single-field and multi-field models

Methodology Details

Problem Definition

Finding the bounce solution satisfying the equations of motion and boundary conditions at infinity: $-\nabla^2\phi_i + \frac{\partial V}{\partial \phi_i} = 0$$\lim_{|x|\to\infty} \phi_i(x) = 0$

where $T[\phi] > 0, V[\phi] < 0$ (non-trivial solution) is required.

Core Method Architecture

CGM Simplification Problem

Transforms bounce solution calculation into a constrained optimization problem: minimizing kinetic energy $T$ under the constraint of fixed negative potential energy $V < 0$.

Gradient Flow Equation

Proposes the following flow equation: $\frac{\partial}{\partial \tau}\phi_i(r,\tau) = \nabla^2\phi_i - \lambda[\phi]\frac{\partial V(\phi)}{\partial \phi_i}$

where the Lagrange multiplier is: $\lambda[\phi] = \frac{\sum_i \int_0^\infty dr r^{d-1} \frac{\partial V(\phi)}{\partial \phi_i}\nabla^2\phi_i}{\sum_i \int_0^\infty dr r^{d-1} \left(\frac{\partial V(\phi)}{\partial \phi_i}\right)^2}$

Scale Transformation Recovery

The bounce solution is obtained through scale transformation: $\phi_B(r) = \phi(\lambda^{-1/2}r)$

Technical Innovations

1. Constraint Preservation: The flow equation ensures $\frac{d}{d\tau}V[\phi] = 0$, keeping potential energy constant during evolution
2. Monotonic Decrease: Kinetic energy monotonically decreases $\frac{d}{d\tau}T[\phi] \leq 0$, guaranteeing convergence to minimum
3. Avoids Saddle Points: Bypasses direct treatment of unstable directions by fixing potential energy
4. Theoretical Guarantees: CGM theorem ensures existence and optimality of solutions

Experimental Setup

Test Models

Single-field Model

$V(\phi) = \frac{1}{2}\phi^2 - \frac{1}{3}\phi^3$ tested in 4-dimensional Euclidean space.

Two-field Model

$V = (\phi_1^2 + 5\phi_2^2)(5(\phi_1-1)^2 + (\phi_2-1)^2) + c\left(\frac{1}{4}\phi_2^4 - \frac{1}{3}\phi_2^3\right)$ tested with parameters $c = 2$ and $c = 80$.

Evaluation Methods

Comparison with existing tool CosmoTransitions to verify solution accuracy.

Initial Condition Testing

Robustness of convergence verified using different initial configurations:

• Piecewise linear functions
• Initial conditions with varying amplitudes and support intervals

Experimental Results

Main Results

Convergence Verification

• Single-field Case: Figures 1-3 show convergence processes under different initial conditions, all converging to stable solutions
• Solution Consistency: Final configurations from different initial conditions are related through appropriate scale transformations

Accuracy Verification

• Single-field Model: Figure 4 shows high consistency between this method and CosmoTransitions
• Two-field Model: Figures 5-8 show results for $c=2$ and $c=80$ cases, matching CosmoTransitions results in both $r-\phi$ and $\phi_1-\phi_2$ planes

Performance Characteristics

• Computational Speed: O(0.1) seconds for computation
• Accuracy: O(0.1)% computational precision
• Applicable Range: Supports 1-8 scalar field models
• Efficiency Advantage: Faster than existing tools like CosmoTransitions, AnyBubble, BubbleProfiler

Robustness Analysis

Testing with different initial conditions demonstrates that as long as $V[\phi]|_{\tau=0} < 0$ is satisfied, the method converges to the correct solution, demonstrating good robustness.

Related Work

Main Research Directions

1. Modified Gradient Flow Methods: Early work by Claudson et al.
2. Modified Action Methods: Kusenko et al.'s approach making bounce solutions local minima
3. Numerical Algorithms: Various improved numerical solution techniques
4. Approximation Methods: Thin-wall approximation, multi-field polygonal bounces
5. Machine Learning: Neural networks for solving differential equations

Advantages of This Work

Compared to existing methods, this approach offers theoretical guarantees, computational simplicity, and stable convergence, particularly superior performance in multi-field cases.

Conclusions and Discussion

Main Conclusions

1. Proposes a new gradient flow equation based on CGM simplification
2. Theoretically guarantees method convergence and optimality
3. Numerical verification shows the method is accurate, efficient, and robust
4. The developed numerical package significantly improves computational efficiency

Limitations

1. Dimensional Constraint: Requires spatial dimension $d > 2$
2. Potential Function Constraints: Requires satisfaction of specific potential function conditions
3. Spherical Symmetry Assumption: Utilizes the spherical symmetry property of bounce solutions

Future Directions

1. Extension to more general potential function forms
2. Consideration of gravitational effects
3. Further optimization of numerical implementation

In-Depth Evaluation

Strengths

1. Solid Theoretical Foundation: Based on classical CGM theory with rigorous mathematical guarantees
2. Elegant Methodology: Transforms complex saddle point problems into constrained minimization problems
3. Excellent Numerical Performance: Significant improvements in computational speed and accuracy
4. Strong Robustness: Insensitive to initial conditions with stable convergence
5. High Practical Value: Open-source code package facilitates application

Weaknesses

1. Limited Theoretical Innovation: Primarily a numerical implementation improvement of CGM theory
2. Relatively Simple Test Cases: Although multi-field cases are included, potential function forms are relatively standard
3. Insufficient Physical Insight: Focuses more on numerical methods than physical mechanisms

Impact

1. Academic Contribution: Provides more efficient tools for false vacuum decay calculations
2. Practical Value: Significantly improves computational efficiency for related calculations
3. Reproducibility: Open-source code facilitates verification and application

Application Scenarios

1. Particle Physics: Phase transition research in the Standard Model and its extensions
2. Cosmology: Early universe phase transitions and bubble nucleation processes
3. Field Theory Research: Non-perturbative phenomena in general scalar field theories

References

The paper cites 37 important references, encompassing classical work on false vacuum decay theory (Coleman, CGM, etc.) and recent advances (work by Chigusa et al.), as well as related numerical methods and software tools, demonstrating comprehensive and in-depth understanding of the field.