Non-uniqueness for the Complex Ginzburg-Landau Equation
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.
Ada: I'm Ada.
Emmy: And I'm Emmy, here with Kolya.
Ada: We're looking at the paper: "Non-uniqueness for the Complex Ginzburg-Landau Equation".
Emmy: Here's the gist of the paper. Singularities of the three-dimensional cubic complex Ginzburg–Landau equation developing from smooth initial data can induce non-uniqueness of solutions after the time of blowup The work establishes this by showing that singularities lead to non-uniqueness because they are related to an unstable eigenvalue of the linearization around the forward self-similar profile generated by the singularity This finding is motivated by similar open problems in Navier–Stokes theory regarding uniqueness for sufficiently regular initial conditions The paper explores the Cauchy problem for the complex Ginzburg–Landau equation (1) and its relation to the Navier–Stokes equation (2), noting that non-uniqueness in CGL is relevant because it shares properties like an energy inequality and scaling symmetry with Navier–Stokes The authors propose a three-step approach to establish the result: (i) establishing the existence of self-similar singularities, (ii) showing that taking the blowup profile as initial data leads to non-uniqueness, and (iii) showing that this situation can be preserved for a suitable truncation leading to self-similar blowup Existence of Unstable Eigenvalue The core of the proof involves establishing the existence of an unstable eigenvalue for the linearized operator LQˆ, which is achieved through a computer-assisted proof This is done by constructing backward and forward self-similar solutions and showing that they match at an intermediate point to form a global solution The existence of this eigenvalue is verified by showing that the function H(λ) has a zero using a winding number argument, which confirms the eigenvalue is unstable because its real part is positive Non-uniqueness Mechanism The non-uniqueness scenario arises when an unstable eigenvalue λ of LQˆ exists with Re λ > 0 This instability allows for the separation of solutions along the unstable manifold, leading to a second solution v defined for some time interval (T > 0) that coincides with the forward self-similar solution at t = -T The existence of this second solution is confirmed by showing that the operator T maps a ball into itself and has a unique fixed point using the Banach fixed-point theorem Solution Construction and Estimates The paper constructs two distinct radial solutions, u1 and u2, which share the same initial data at t = -T but are different for t ∈ (0, T] These solutions are shown to be locally suitable weak solutions of the CGL equation with an equality in the energy inequality The difference between them is bounded in Hα(R 3) for any α ∈
0, 1/2) Linearized Solution Analysis The linearized solution at infinity is studied by analyzing the operator LQˆ, which is shown to generate a hyperbolic rescaled semigroup after rescaling This allows for the decomposition of the space into stable and unstable subspaces, leading to exponential stability estimates for solutions in these subspaces The analysis of the linearized equation at zero is also performed by splitting the interval [0, ξ1: and using a Taylor expansion at zero Forward Solution Analysis The forward self-similar solution is found by solving the associated ODE (44) via variation of parameters, leading to a fixed-point operator Tˆ The existence and uniqueness of this fixed point are established by showing that Tˆ is a contraction in a weighted function space under appropriate conditions on the parameters Linearized Solution at Infinity The linearized solution at infinity is analyzed by diagonalizing the operator LQˆ and splitting the system into two scalar ODEs The solutions to these scalar equations are found using variation of parameters, leading to a fixed-point equation for Z(ξ) which yields solutions to (55) Linearized Solution at Zero The linearized solution at zero is analyzed by splitting the interval
0, ξ1: and using a Taylor expansion at zero for the first interval and a rigorous numerical integrator for the second interval The resulting system of first-order equations is then implemented in CAPD Implementation Details The computer-assisted proofs rely on the CAPD library for rigorous numerical integration and the Arb library for other computations The implementation details include using a finite-difference scheme to compute an initial approximation of an unstable eigenvalue of LQˆ The full code is available in the repository
6: The authors take full responsibility for the content of both the manuscript and the code Asymptotic Behavior The asymptotic behavior of the backward self-similar solution at infinity is determined by a lemma showing that Q(ξ) = c−a pQ,0ξ−2a + O(ξ−1σ−2) The leading asymptotic behavior of the forward self-similar solution is also derived using variation of parameters to construct Qˆ(ξ) = ˆγ1Pˆ(ξ) + ˆγ2Eˆ(ξ) + Pˆ(ξ)IE(Q, ξ ̂) + Eˆ(ξ)IP (Q, ξ ̂) Numerical Computation of Eigenvalue An initial approximation of an unstable eigenvalue of LQˆ is computed using a finite-difference scheme on a uniform grid with Neumann and Dirichlet conditions The eigenvalues are then computed using the ARPACK sparse eigenvalue solver The winding number argument is used to prove the existence of a unique zero inside a square contour around the approximate solution Asymptotic Behavior of Backward Solution The asymptotic behavior of the backward self-similar solution at infinity is described by Lemma B.1, which states that Q(ξ) = c−a pQ,0ξ−2a + O(ξ−1σ−2) This behavior is crucial for computing the forward solution and confirms that the leading asymptotic behavior of Q is not given by the term γP alone The integral IE,∞ is bounded using Lemma B.2 which provides bounds for IE,∞ in terms of γ and other quantities Enclosures for Fixed Point Refined enclosures of the fixed point Z(ξ) are computed by isolating the leading term of the integrand and integrating it explicitly to bound IEˆ(ξ) The bounds for QˆRe ˆγ2 and Qˆ Im ˆγ2 are derived using Lemma 5.8 which provides bounds in terms of the norm of Q̂ Linearized Solution at Infinity The linearized solution at infinity is analyzed by diagonalizing the operator and treating the right-hand side as a perturbation The resulting scalar ODEs are solved using variation of parameters to find solutions P1(ξ), P2(ξ), E1(ξ), and E2(ξ) The Wronskians are used to simplify the matching condition for finding solutions to (55) Linearized Solution at Zero The linearized solution at zero is studied by splitting the interval
0, ξ1: and using a Taylor expansion at zero for the first interval and a rigorous numerical integrator for the second interval The resulting 12-dimensional scalar system of first-order equations is then implemented in CAPD Linearized Solution at Infinity The linearized solution at infinity is studied by diagonalizing the operator and treating the right-hand side as a perturbation The resulting scalar ODEs are solved using variation of parameters to find solutions P1(ξ), P2(ξ), E1(ξ), and E2(ξ) Implementation Details The full code on which the computer-assisted parts of the proofs are based is available at https://github.com/Joel-Dahne/CGL2.jl The code uses Julia for most implementations and C++ for CAPD related parts The rigorous parts utilize the CAPD library and the Arb/FLINT library The authors take full responsibility for the content of both the manuscript and the code Asymptotic Behavior of Backward Solution The asymptotic behavior of the backward self-similar solution at infinity is described by Lemma B.1, which states that Q(ξ) = c−a pQ,0ξ−2a + O(ξ−1σ−2) This behavior is crucial for computing the forward solution and confirms that the leading asymptotic behavior of Q is not given by the term γP alone Linearized Solution at Infinity The linearized solution at infinity is analyzed by diagonalizing the operator and treating the right-hand side as a perturbation The resulting scalar ODEs are solved using variation of parameters to find solutions P1(ξ), P2(ξ), E1(ξ), and E2(ξ) Asymptotic Behavior of Backward Solution The asymptotic behavior of the backward self-similar solution at infinity is described by Lemma B.
Paper summary: Ada: So we're looking at this paper, "Non-uniqueness for the Complex Ginzburg-Landau Equation," and what it claims is that singularities in three dimensions for this cubic equation can cause different solutions to emerge after they blow up in time.
Emmy: Right. The core thesis of this paper is that these singularities lead to non-uniqueness because they are connected to an unstable eigenvalue found when linearizing around the forward self-similar profile generated by that singularity.
Kolya: So, what's the big deal here? Why should we care about non-uniqueness in a complex Ginzburg-Landau equation when we're already dealing with things like Navier-Stokes?
Ada: Well, it matters because the paper is motivated by similarities between these two equations. They share things like an energy inequality and a scaling symmetry. The authors suggest that this situation mirrors open problems in Navier-Stokes theory concerning uniqueness for initial conditions that are just regular enough.
Emmy: They approach this problem in three steps: first, they establish the existence of these self-similar singularities, then they show taking those blowup profiles as starting data creates non-uniqueness, and finally, they look at how this situation persists when you use a suitable truncation to get a self-similar blowup.
Kolya: That sounds like a pretty deep dive into constructing these solutions. What's the most technical part they focus on?
Ada: The central part is proving the existence of an unstable eigenvalue for the linearized operator, which they manage to establish using a computer-assisted proof. They do this by constructing both backward and forward self-similar solutions and showing they match at some intermediate point to form a global solution.
Emmy: And they verify that this eigenvalue is indeed unstable by showing that some function H has a zero using a winding number argument, which confirms the real part of the eigenvalue is positive.
Kolya: Okay, so an unstable eigenvalue with a positive real part means solutions can split along the unstable manifold. What does that actually mean for finding different solutions?
Ada: That instability allows you to separate solutions along that unstable manifold, meaning there's a second solution defined for some time interval T greater than zero that matches the forward self-similar solution at time negative T. They confirm this second solution exists by showing the operator T maps a ball into itself and has a unique fixed point using the Banach fixed-point theorem.
Emmy: The paper then constructs two distinct radial solutions, u one and u two, which start with the same initial data at time negative T but are different for all times between zero and T.
Kolya: And what's the bound on how different these two solutions are? Is it small?
Ada: They show that the difference between them is bounded in H alpha of R three for any alpha less than one over two. This means they are locally suitable weak solutions, and they satisfy an equality in the energy inequality.
Emmy: Looking at the linearized solution at infinity, they analyze the operator LQ hat and show it generates a hyperbolic rescaled semigroup after rescaling. This lets them split the space into stable and unstable subspaces, giving exponential stability estimates for solutions in those subspaces.
Kolya: What about the other side? How do they handle the behavior near zero? That's often tricky in these kinds of problems.
Ada: They also study the linearized solution at zero by splitting the interval
zero, xi one: . For that first part, they use a Taylor expansion at zero, and for the second part, they use a rigorous numerical integrator. This results in a twelve-dimensional system of first-order equations that they implement in CAPD.
Emmy: And to get those solutions to infinity, they diagonalize the operator LQ hat and split it into two scalar ordinary differential equations. Solving those scalar equations using variation of parameters leads to a fixed point equation for Z xi, which yields solutions to the system of five-dimensional equations.
Kolya: So you have these intricate numerical tools and analytical methods working together to show this non-uniqueness happens? What's on the horizon for this research?
Ada: The authors state that they left step three—showing that this situation can be preserved for a suitable truncation leading to self-similar blowup—for future work, but they feel confident it's reachable with existing methods.
Emmy: They also discuss the nature of the non-uniqueness itself. They suggest it seems hard to find some kind of selection principle that would pick out "the right solution."
Kolya: So what does this say about the model itself? The paper hints at something bigger than just solving this specific PDE.
Ada: It suggests that the model might be incomplete. They imply that to figure out what happens when uniqueness is lost, you have to look into the "next layer" of physics, which they call the "effective theory" for the complex Ginzburg-Landau equation itself being insufficient.
Emmy: That connects back to their motivation with Navier-Stokes because for that equation, it's currently unclear if a similar situation could occur.
Kolya: It sounds like they are pointing toward needing a more fundamental model to fully understand when uniqueness fails in these types of systems.
Ada: So, the paper "Non-uniqueness for the Complex Ginzburg-Landau Equation" shows that singularities can lead to different solutions because of an unstable eigenvalue, and it points toward the need for a deeper physical understanding of these models.
Conclusion: Ada: So we've been looking at how singularities in this complex Ginzburg Landau equation cause solutions to become non-unique after they blow up in time.
Emmy: It really hinges on these instabilities found when you linearize around the forward self-similar profile that those singularities create.
Kolya: It sounds like a pretty deep dive into the math, but what does this actually mean for us who aren't doing complex field theory?
Ada: Well, it connects to things like Navier Stokes where we already worry about whether solutions are unique when initial conditions are just regular enough.
Emmy: The paper is essentially showing that these specific types of singularities introduce a mathematical pathway for two different solutions to exist starting from the same place.
Kolya: So the numbers they're using, they're showing that there's an unstable eigenvalue with a positive real part, which lets them separate those solutions.
Ada: Exactly. That instability creates this direction in the solution space where you can branch off into a second solution defined for some time interval.
Emmy: The authors construct two distinct radial solutions that start with identical initial data but diverge after some time T because of this instability.
Kolya: And they show that the difference between those two solutions is bounded, specifically in a specific Sobolev space, H alpha of R three where alpha is less than one over two.
Ada: That bound is important because it means these aren't just wildly different; they are locally suitable weak solutions that still satisfy the energy inequality exactly.
Emmy: They also break down the linearized operator at infinity by diagonalizing it, which splits the space into stable and unstable parts, giving us some exponential stability estimates for those subspaces.
Kolya: It seems like they're really digging into how these mathematical structures lead to physical ambiguity when things go singular.
Ada: That's what they are doing, trying to find the source of that ambiguity in this complex system itself.
Emmy: They also hint that there might be a limitation in the effective theory for this equation, suggesting we need a more fundamental model to fully understand these non-unique situations.
Kolya: So while they solved the uniqueness problem for a specific part, they're opening up questions about what else is missing in our current understanding of this model.
math.AP
Submitted: 2026-09-08
Updated: 2026-11-09
Comments: 63 pages, 4 figures
Code: https://github.com/Joel-Dahne/CGL2.jl
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 67/100
The gist: Singularities of the three-dimensional cubic complex Ginzburg–Landau equation developing from smooth initial data can induce non-uniqueness of solutions after the time of blowup The work…
Key concepts
- Unstable Eigenvalue
- This is a specific mathematical property of the linearized operator (LQˆ) around a self-similar solution. If its real part is positive (Re λ > 0), it indicates instability. This instability means small perturbations grow exponentially, which is the key mechanism that allows two different solutions to exist from the same starting point.
- Self-similar Singularities
- These are specific types of solutions where the solution's shape evolves in a predictable way as time approaches a blowup. The paper establishes their existence and uses them as a foundation. By analyzing these profiles, the authors can understand how singularities arise in the complex Ginzburg–Landau equation.
- Linearized Operator (LQˆ)
- This operator describes how small changes propagate around a known solution (the forward self-similar profile). Analyzing its eigenvalues helps determine the stability of that solution. Finding an unstable eigenvalue is crucial because it provides the mathematical pathway for constructing a second, distinct solution.
- Non-uniqueness Mechanism
- The non-uniqueness arises when the unstable eigenvalue exists with a positive real part. This instability permits solutions to diverge along an 'unstable manifold.' This allows one to construct a second solution that differs from the main forward solution after some time interval.
Terminology
Summary
Singularities of the three-dimensional cubic complex Ginzburg–Landau equation developing from smooth initial data can induce non-uniqueness of solutions after the time of blowup The work establishes this by showing that singularities lead to non-uniqueness because they are related to an unstable eigenvalue of the linearization around the forward self-similar profile generated by the singularity This finding is motivated by similar open problems in Navier–Stokes theory regarding uniqueness for sufficiently regular initial conditions
Motivation and Approach
The paper explores the Cauchy problem for the complex Ginzburg–Landau equation (1) and its relation to the Navier–Stokes equation (2), noting that non-uniqueness in CGL is relevant because it shares properties like an energy inequality and scaling symmetry with Navier–Stokes The authors propose a three-step approach to establish the result: (i) establishing the existence of self-similar singularities, (ii) showing that taking the blowup profile as initial data leads to non-uniqueness, and (iii) showing that this situation can be preserved for a suitable truncation leading to self-similar blowup
Existence of Unstable Eigenvalue
The core of the proof involves establishing the existence of an unstable eigenvalue for the linearized operator LQˆ, which is achieved through a computer-assisted proof This is done by constructing backward and forward self-similar solutions and showing that they match at an intermediate point to form a global solution The existence of this eigenvalue is verified by showing that the function H(λ) has a zero using a winding number argument, which confirms the eigenvalue is unstable because its real part is positive
Non-uniqueness Mechanism
The non-uniqueness scenario arises when an unstable eigenvalue λ of LQˆ exists with Re λ > 0 This instability allows for the separation of solutions along the unstable manifold, leading to a second solution v defined for some time interval (T > 0) that coincides with the forward self-similar solution at t = -T The existence of this second solution is confirmed by showing that the operator T maps a ball into itself and has a unique fixed point using the Banach fixed-point theorem
Solution Construction and Estimates
The paper constructs two distinct radial solutions, u1 and u2, which share the same initial data at t = -T but are different for t ∈ (0, T] These solutions are shown to be locally suitable weak solutions of the CGL equation with an equality in the energy inequality The difference between them is bounded in Hα(R 3) for any α ∈ [0, 1/2)
Linearized Solution Analysis
The linearized solution at infinity is studied by analyzing the operator LQˆ, which is shown to generate a hyperbolic rescaled semigroup after rescaling This allows for the decomposition of the space into stable and unstable subspaces, leading to exponential stability estimates for solutions in these subspaces The analysis of the linearized equation at zero is also performed by splitting the interval [0, ξ1] and using a Taylor expansion at zero
Forward Solution Analysis
The forward self-similar solution is found by solving the associated ODE (44) via variation of parameters, leading to a fixed-point operator Tˆ The existence and uniqueness of this fixed point are established by showing that Tˆ is a contraction in a weighted function space under appropriate conditions on the parameters
Linearized Solution at Infinity
The linearized solution at infinity is analyzed by diagonalizing the operator LQˆ and splitting the system into two scalar ODEs The solutions to these scalar equations are found using variation of parameters, leading to a fixed-point equation for Z(ξ) which yields solutions to (55)
Linearized Solution at Zero
The linearized solution at zero is analyzed by splitting the interval [0, ξ1] and using a Taylor expansion at zero for the first interval and a rigorous numerical integrator for the second interval The resulting system of first-order equations is then implemented in CAPD
Implementation Details
The computer-assisted proofs rely on the CAPD library for rigorous numerical integration and the Arb library for other computations The implementation details include using a finite-difference scheme to compute an initial approximation of an unstable eigenvalue of LQˆ The full code is available in the repository [6] The authors take full responsibility for the content of both the manuscript and the code
Asymptotic Behavior
The asymptotic behavior of the backward self-similar solution at infinity is determined by a lemma showing that Q(ξ) = c−a pQ,0ξ−2a + O(ξ−1σ−2) The leading asymptotic behavior of the forward self-similar solution is also derived using variation of parameters to construct Qˆ(ξ) = ˆγ1Pˆ(ξ) + ˆγ2Eˆ(ξ) + Pˆ(ξ)IE(Q, ξ ̂) + Eˆ(ξ)IP (Q, ξ ̂)
Numerical Computation of Eigenvalue
An initial approximation of an unstable eigenvalue of LQˆ is computed using a finite-difference scheme on a uniform grid with Neumann and Dirichlet conditions The eigenvalues are then computed using the ARPACK sparse eigenvalue solver The winding number argument is used to prove the existence of a unique zero inside a square contour around the approximate solution
Asymptotic Behavior of Backward Solution
The asymptotic behavior of the backward self-similar solution at infinity is described by Lemma B.1, which states that Q(ξ) = c−a pQ,0ξ−2a + O(ξ−1σ−2) This behavior is crucial for computing the forward solution and confirms that the leading asymptotic behavior of Q is not given by the term γP alone The integral IE,∞ is bounded using Lemma B. p. 0 Contents
1 Introduction 1
2 Non-uniqueness from an unstable eigenvalue 8
3 Backward, forward and linearized solutions 20
4 Existence of an unstable eigenvalue 23
5 Forward solution at infinity 28
6 Forward solution at zero 40
7 Linearized solution at infinity 40
8 Linearized solution at zero 50
9 Implementation details 54
A Finite-difference approximations of eigenvalues 55
B Asymptotic behavior of the backward self-similar solution 56
1 Introduction
We consider the Cauchy problem for the complex Ginzburg–Landau equation for functions u: R
3 ×
(0, ∞) → C
i
∂u
∂t + (1 − iϵ)∆u + (1 + iδ)|u|
2u = 0 in R
3 × (0, ∞),
u|t=0 = u0,
(1)
where the initial condition u0 belongs to a suitable class of functions to be specified later. We
assume ϵ > 0 and δ ≥ 0 are given parameters.Where the paper says it
Enclosures for Fixed Point
Refined enclosures of the fixed point Z(ξ) are computed by isolating the leading term of the integrand and integrating it explicitly to bound IEˆ(ξ) The bounds for QˆRe ˆγ2 and Qˆ Im ˆγ2 are derived using Lemma 5.8 which provides bounds in terms of the norm of Q̂
The linearized solution at infinity is analyzed by diagonalizing the operator and treating the right-hand side as a perturbation The resulting scalar ODEs are solved using variation of parameters to find solutions P1(ξ), P2(ξ), E1(ξ), and E2(ξ) The Wronskians are used to simplify the matching condition for finding solutions to (55)
The linearized solution at zero is studied by splitting the interval [0, ξ1] and using a Taylor expansion at zero for the first interval and a rigorous numerical integrator for the second interval The resulting 12-dimensional scalar system of first-order equations is then implemented in CAPD
The linearized solution at infinity is studied by diagonalizing the operator and treating the right-hand side as a perturbation The resulting scalar ODEs are solved using variation of parameters to find solutions P1(ξ), P2(ξ), E1(ξ), and E2(ξ)
The full code on which the computer-assisted parts of the proofs are based is available at https://github.com/Joel-Dahne/CGL2. p. 0 Non-uniqueness for the Complex Ginzburg–Landau Equation
Joel Dahne, Vladim´ır Sver´ak ˇ
Abstract
We show that singularities of the three-dimensional cubic complex Ginzburg–Landau equation developing from smooth initial data can induce non-uniqueness of solutions after the time
of blowup. The solutions we study start as backward self-similar solutions, and the source of
the non-uniqueness is an unstable eigenvalue of the linearization around the forward self-similar
profile generated by the singularity.Where the paper says it
The asymptotic behavior of the backward self-similar solution at infinity is described by Lemma B.1, which states that Q(ξ) = c−a pQ,0ξ−2a + O(ξ−1σ−2) This behavior is crucial for computing the forward solution and confirms that the leading asymptotic behavior of Q is not given by the term γP alone
The linearized solution at infinity is analyzed by diagonalizing the operator and treating the right-hand side as a perturbation The resulting scalar ODEs are solved using variation of parameters to find solutions P1(ξ), P2(ξ), E1(ξ), and E2(ξ)
The asymptotic behavior of the backward self-similar solution at infinity is described by Lemma B.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved ability to analyze complex nonlinear evolution equations by leveraging backward self-similar solutions to establish non-uniqueness from unstable eigenvalues. This allows for rigorous proof of solution non-uniqueness in cases related to Navier–Stokes equation modeling, as shown by
We show that singularities of the three-dimensional cubic complex Ginzburg–Landau equation developing from smooth initial data can induce non-uniqueness of solutions after the time of blowup.
-
Enhanced capability in numerical analysis for finding unstable eigenvalues using computer-assisted proofs. The system can compute an approximate eigenvalue using a finite-difference scheme and then employ the interval Newton method, as described in
To compute an initial approximation of an unstable eigenvalue of LQˆ, we make use of a finitedifference scheme.
-
Advanced techniques for constructing solutions to nonlinear evolution equations by utilizing a Duhamel-type argument within a weighted Banach space. The system can construct perturbed solutions using the operator T which is shown to be a strict contraction in specific balls, as detailed in
By the Banach fixed-point theorem, T has a unique fixed point.
-
Precise mathematical verification of solution properties across different time regimes and spatial domains. The system can demonstrate that solutions u1 and u2 share the same initial data at t = -T and are distinct for t in (0, T], as shown in
Then u1 and u2 are two different radial solutions to the CGL equation having the same initial data at t = -T.
-
Robust analysis of asymptotic behavior of singular solutions at spatial infinity. The system can determine the leading asymptotic behavior of a backward self-similar solution by showing that
Q(ξ) = c−a pQ,0ξ−2a + O(ξ−1σ−2),
which is critical for understanding the structure of non-uniqueness in the limit. -
Automated derivation of bounds on linear and nonlinear operators associated with linearized equations at infinity. The system can compute operator norms like K1(xi)∞,j and K2(xi)∞,j using Lemma 7.2 to control the growth of solutions as spatial frequency increases.
Abstract
We show that singularities of the three-dimensional cubic complex Ginzburg-Landau equation developing from smooth initial data can induce non-uniqueness of solutions after the time of blowup. The solutions we study start as backward self-similar solutions, and the source of the non-uniqueness is an unstable eigenvalue of the linearization around the forward self-similar profile generated by the singularity. We establish the existence of this eigenvalue, and hence the non-uniqueness, by a computer-assisted proof for specific values of the parameters in the equation. The work is in part motivated by similarities to conjectured behavior for the Navier-Stokes equation, with which the complex Ginzburg-Landau equation shares several important properties, such as an energy inequality and the scaling symmetry.
Sources
- Self-Similar Singular Solutions to the Nonlinear Schr\"odinger and the Complex Ginzburg-Landau Equations
- On the non-uniqueness of solutions of the axi-symmetric swirl-free Navier-Stokes equations, I
Related papers
- Regularity of critical immersions for scale-invariant curvature energies in even dimensions
- C 1, alpha Regularity of W 2,2 Q-Valued Maps with Small Laplacian
- Cumulative angular cancellation in Maz'ya's endpoint inequalities
- Strictly positive optimal Hardy weights on manifolds and graphs via Green potential
- A fast dynamo with zero topological entropy
- A counterexample to fixed-profile soliton resolution for the three-dimensional quintic wave equation