Existence of Solutions of the third term of the Connaughton-Newell Model with a source term
Anh Việt Nguyễn · 2026
Read Existence of Solutions of the third term of the Connaughton-Newell Model with a source term by Anh Việt Nguyễn (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-14162.txt.
Introduction
Nguyen studies the Connaughton-Newell equation, a coagulation-fragmentation model that approximates three-wave kinetic equations. The model has three nonlinear operators, and this work targets the third one. With a constant interaction kernel and a well-behaved source term, the author proves that the third operator has a solution. The result adds an existence guarantee for one part of this fully nonlinear model. The proof stays within the model's own setting.
Catalog record
| CREATOR | Anh Việt Nguyễn |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 13873 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.14162
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Existence of Solutions of the third term of the Connaughton-Newell Model with a source term Anh Việt Nguyễn Abstract . The Connaughton-Newell equation is an approximation of three-wave kinetic equations using a fully non-linear coagulation-fragmentation model. This equation consists of three non-linear operators. In this paper, we proved that assuming a constant interaction kernel and a well-behaved source term, the third operator of the Connaughton-Newell equation has a solution. 1. Introduction The theory of wave turbulence provides a fundamental framework for understanding a wide range of physical and mechanical phenomena. Some notable examples are waves due to rotation, Alfvén wave turbulence in solar wind, and waves in plasma fusion devices. the theory of wave turbulence, first formulated by Peiers´ [ 1 ] , was further developed by Benney & Saffman [ 2 ] , Zakharov & Filonenko [ 3 ] , Benney & Newell [ 4 ] , and most notably, Hasselmann [ 5 , 6 ] . These works led to the formulation of the three- and four-wave kinetic equations, which describe the energy distribution among weakly interacting waves. The three-wave kinetic equation is written \partial_{t}f(t,p):=\int\int_{\mathbb{R}^{2d}}[\mathrm{R}_{p,p_{1},p_{2}}[f]-\mathrm{R}_{p_{1},p,p_{2}}[f]-\mathrm{R}_{p_{2},p,p_{1}}[f]]dp_{1}dp_{2},\quad f(0,p)=f_{0}(p), where, \mathrm{R}_{p,p_{1},p_{2}}[f] is defined as \mathrm{R}_{p,p_{1},p_{2}}[f]:=|V_{p,p_{1},p_{2}}|^{2}\delta(p-p_{1}-p_{2})\delta(\omega(p)-\omega(p_{1})-\omega(p_{2}))(f(p_{1})f(p_{2})-ff(p_{1})-ff(p_{2})). f(t,p) represents the wave density at wave number p\in\mathbb{R}^{d} with initial condition f_{0}(p) , and \omega(p) represents the dispersion relation of the waves. For further understanding, we direct readers to the comprehensive work in [ 7 , 8 , 9 ] . In a closely related setting, the coagulation and fragmentation kinetic equations also appear in a variety of physical and mechanical applications. These include: the formation of aerosols, polymers, celestial bodies on an astronomical scale, and colloidal aggregates [ 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 ] . These equations describe the coagulating or fragmentation behavior of a system of particles under some interaction or external force. For the pure coagulation model, the Smoluchowski equation, where Well-Posedness was shown by Shirvani and Roessel [ 22 ] , serves as a classical model for particulate processes. In the study of these equations, the main focus is on the particle density function, usually denoted as n_{\omega}(t) where \omega\in\mathbb{N} denotes the size of particles at time t\in[0,+\infty) . Surprisingly, there is a conceptual similarity between the three-wave kinetic equations and the coagulation-fragmentation kinetic equations. In particular, the transfer of energy between scales in wave turbulence is similar to the transfer of mass between clusters in the coagulation process. In 2010, Connaughtin and Newell proposed an approximation of the three-wave kinetic equations using a non-linear coagulation-fragmentation model [ 23 ] . A numerical scheme for this model was provided by Das and Tran [ 24 ] , and its analysis was provided by Wang and Tran [ 25 ] . The Connaughton-Newell equation expresses both the coagulation and fragmentation terms in a nonlinear manner. The equation read \frac{\partial N_{\omega}}{\partial t}:=\mathcal{Q}[N_{\omega}](t),\ \omega\in\mathbb{R^{+}},\ N_{\omega}(0)=N_{\omega}^{in}, where the operator \mathcal{Q}[N_{\omega}](t) is written as \mathcal{Q}[N_{\omega}](t):=S_{1}[N_{\omega}]+S_{2}[N_{\omega}]+S_{3}[N_{\omega},], and S_{1}(N_{\omega}):=\int_{0}^{\omega}K_{1}(\omega-\mu,\mu)N_{\mu}N_{\omega-\mu}d\mu-\int_{\omega}^{\infty}K_{1}(\mu-\omega,\omega)N_{\omega}N_{\mu-\omega}d\mu-\int_{0}^{\infty}K_{1}(\omega,\mu)N_{\omega}N_{\mu}d\mu, S_{2}(N_{\omega}):=-\int_{0}^{\omega}K_{2}(\mu,\omega-\mu)N_{\omega}N_{\mu}d\mu+\int_{\omega}^{\infty}K_{2}(\omega,\mu-\omega)N_{\mu}N_{\mu-\omega}d\mu+\int_{0}^{\infty}K_{2}(\omega,\mu)N_{\omega+\mu}N_{\mu}d\mu, S_{3}(N_{\omega}):=-\int_{0}^{\omega}K_{3}(\mu,\omega-\mu)N_{\omega}N_{\omega-\mu}d\mu+\int_{\omega}^{\infty}K_{3}(\mu,\mu-\omega)N_{\omega}N_{\mu}d\mu+\int_{0}^{\infty}K_{3}(\mu,\omega)N_{\omega+\mu}N_{\mu}d\mu. Here, N_{\mu} represents the wave frequency spectrum such that \int_{\mu_{1}}^{\mu_{2}}N_{\mu}d\mu is the total wave action in the frequency band [\mu_{1},\mu_{2}] . The homogeneous function K_{i}(\omega,\mu) represents the wave interaction kernels where K_{1} is associated with the transfer of energy as K_{2} and K_{3} are associated with the back-scattering of energy. In this paper, we show that there is a solution to the equation \frac{\partial N_{\omega}}{\partial t}=S_{3}(N_{\omega}) with some additional assumptions. In particular, we look at S_{3}(N_{\omega}) with a non-trivial source term, denoted g_{\omega}(t) , and a constant interaction kernel, denoted K_{3}(\omega,\mu) , which will be regarded as 1. We consider the equation (1) \begin{split}\frac{\partial N_{\omega}}{\partial t}&=-\int_{0}^{\omega}N_{\omega}N_{\omega-\mu}d\mu+\int_{\omega}^{\infty}N_{\omega}N_{\mu}d\mu+\int_{0}^{\infty}N_{\omega+\mu}N_{\mu}d\mu+g_{\omega}(t).\\ \end{split} Equation ( 1 ) can be rewritten as (2) \begin{split}\frac{\partial N_{\omega}}{\partial t}=N_{\omega}\int_{0}^{\infty}N_{\mu}d\mu+\int_{0}^{\infty}N_{\omega+\mu}N_{\mu}d\mu+g_{\omega}(t).\\ \end{split} To account for physical constraints, we impose the following assumptions: (A1) For each \omega\geq 0 , the source term g_{\omega}:[0,T]\to\mathbb{R} belongs to L^{1}_{\mathrm{loc}}[0,T] . (A2) The initial value for N_{\omega} and the source term, g_{\omega} , are nonnegative; that is N_{\omega}(0)\geq 0,\ \text{ for all }\omega\geq 0\quad\text{and}\quad g_{\omega}(t)\geq 0,\ \text{ for all }(t,\omega)\in[0,T]\times(0,\infty).\\ (A3) The initial value is integrable with respect to \omega \int_{0}^{\infty}N_{\omega}(0)\,d\omega<\infty.\\ (A4) The total source term is locally integrable in time \int_{0}^{\infty}g_{\omega}(t)\,d\omega\in L^{1}_{\mathrm{loc}}[0,T]. With these assumptions, we have the following result. Theorem 1 . Assuming that (A1) – (A4) hold, then equation ( 2 ) admits at least one solution. The proof will be provided in the next section. Acknowledgments. The author is thankful for Minh-Binh Tran’s valuable guidance throughout this work. The author also thanks Jennifer Mackenzie for helping to edit this paper. 2. Proof of Theorem 1 We have the following equation (3) \begin{split}N^{\prime}=2N^{2}+g(t),\end{split} obtained by integrating ( 2 ) with respect to \omega , replacing \int_{0}^{\infty}N_{\omega}d\omega with N , and let g(t)=\int_{0}^{\infty}g_{\omega}(t)d\omega . We note that ( 3 ) is a Riccati equation with an arbitrary solution N(t) . Since g(t) is non-negative, N^{\prime}(t) is non-negative. Additionally, given N(t) has a non-negative initial value N(0) , we have the following remark. Remark 2 . The solutions to ( 3 ) are non-negative. Similarly, by replacing \int_{0}^{\infty}N_{\mu}d\mu with N in equation ( 2 ) , we have (4) \begin{split}\frac{\partial N_{\omega}}{\partial t}=N_{\omega}N+\int_{0}^{\infty}N_{\omega+\mu}N_{\mu}d\mu+g_{\omega}(t).\\ \end{split} Since N is non-negative, ( 4 ) has a unique solution given by: (5) \begin{split}N_{\omega}=e^{\psi(t)}N_{\omega}(a)+e^{\psi(t)}\int_{a}^{t}e^{-\psi(t)}\left(\int_{0}^{\infty}N_{\omega+\mu}N_{\mu}d\mu+g_{w}(t)\right)dt,\end{split} where \psi(t)=\int_{a}^{t}Ndt . We aim to show that \int_{0}^{\infty}N_{\mu}d\mu converges uniformly on [0,T] , but first we define the following auxiliary equations: \begin{split}X(a):=\int_{0}^{\infty}N_{\omega}d\omega(a),\quad g(t):=\int_{0}^{\infty}g_{\omega}(t)d\omega,\end{split} \begin{split}d_{\omega}(t):=\sup_{a\leq s\leq t}N_{\omega}(s),\quad m(t):=\sup_{a\leq s\leq t}e^{\psi(s)}\int_{a}^{s}e^{-\psi(\tau)}d\tau,\text{ and}\end{split} p_{\omega}(t):=\sup_{a\leq s\leq t}N_{\omega}(a)e^{\psi(s)}+e^{\psi(s)}\int_{a}^{s}e^{-\psi(\tau)}g_{\omega}(\tau)d\tau. We prove the following lemma. Lemma 3 . Assuming conditions (A1) – (A4) hold, then for every a\geq 0 , there exists b>a such that f_{n}=\int_{0}^{n}N_{\mu}\,d\mu converges uniformly on the interval [a,b] . Proof. We consider p_{\omega}(t) . Since \psi(s) is positive, non-decreasing, and both equations N_{\omega}(a) and g_{\omega}(\tau) are positive, we can replace the e^{\psi(s)}\text{ by }e^{\psi(t)} and e^{-\psi(s)} by 1 . Hence, \begin{split}p_{\omega}(t)\leq N_{\omega}(a)e^{\psi(t)}+e^{\psi(t)}\int_{a}^{t}g_{\omega}(\tau)d\tau.\end{split} Hence, taking the integral on both sides with respect to \omega yields \begin{split}\int_{0}^{\infty}p_{\omega}(t)d\omega\leq&e^{\psi(t)}\int_{0}^{\infty}N_{\omega}d\omega(a)+e^{\psi(t)}\int_{a}^{t}\int_{0}^{\infty}g_{\omega}(\tau)d\omega d\tau\\ =&X(a)e^{\psi(t)}+e^{\psi(t)}\int_{a}^{t}g(\tau)d\tau<\infty.\\ \end{split} Note that the equation ( 5 ) is bounded by \begin{split}p_{\omega}(t)+\sup_{a\leq s\leq t}e^{\psi(s)}\int_{a}^{s}e^{-\psi(\tau)}\int_{0}^{\infty}N_{\mu+\omega}N_{\mu}d\mu d\tau.\end{split} Taking the supremum over t gives \begin{split}d_{\omega}(t)\leq p_{\omega}(t)+m(t)\int_{0}^{\infty}d_{\mu+\omega}d_{\mu}d\mu.\end{split} Next, we define a family of functions e_{\omega}(t) such that e_{0}(t)=d_{0}(t) and \begin{split}e_{\omega}(t)=p_{\omega}(t)+m(t)\int_{0}^{\infty}e_{\mu+\omega}(t)e_{\mu}(t)d\mu.\end{split} We have the following generating functions \begin{split}E(t;z)&=\int_{0}^{\infty}e_{\omega}(t)z^{\omega}d\omega,\text{ and }\\ P(t,z)&=\int_{0}^{\infty}p_{\omega}(t)z^{\omega}d\omega.\\ \end{split} Therefore, \begin{split}E(t,z)&=P(t;z)+mE(t,z)^{2}.\end{split} Thus, \begin{split}E(t,z)=\frac{1\pm\sqrt{1-4Pm}}{2m}.\end{split} We note that m(t)\rightarrow 0 as t\rightarrow a^{+} . Furthermore, since \int_{0}^{\infty}p_{\omega}(t)d\omega is finite, there is b close to a such that \int_{0}^{\infty}p_{\omega}(t)d\omega\leq\frac{1}{4m(t)} for all t\in[a,b] . Then, if |z|\leq 1 , we have: |4m(t)P(t,z)|\leq 4m(t)\int_{0}^{\infty}p_{\omega}(t)(|z|^{\omega})d\omega\leq 4m(t)\int_{0}^{\infty}p_{\omega}(t)d\omega\leq 1. Hence, E(t,1) admits real solutions for all t\in[a,b] . That is: \int_{0}^{\infty}e_{\omega}(t)d\omega<\infty,\text{ for all }t\in[a,b]. We also have \int_{0}^{\tau}N_{\omega}d\omega\leq\int_{0}^{\tau}d_{\omega}d\omega\leq\int_{0}^{\infty}e_{\omega}d\omega\leq\infty,\text{ for all $\tau\in\mathbb{R}$}. Let I_{i}=[i-1,i)\text{ for }i\in\mathbb{N} . We define a sequence of function \{\int_{I_{i}}N_{\omega}\}_{i} and \{\int_{I_{i}}p_{\omega}\}_{i} . Hence, for each i\in I , we have \int_{I_{i}}N_{\omega}\ d\omega\leq\int_{I_{i}}p_{\omega}d\omega\ \leq\int_{I_{i}}e_{\omega}d\omega(b). Therefore, by the Weierstrass M-test, \int_{0}^{\infty}N_{\omega}d\omega converges uniformly. ∎ From this lemma, we know that \int_{0}^{\infty}N_{\omega}d\omega converges uniformly in [0,b] for some b>0 . We let M(t)=\int_{0}^{\infty}N_{\omega}d\omega . We obtain the following result. Lemma 4 . Assuming conditions (A1) – (A4) hold. Then there exists b>0 such that M(t)=N(t),\quad\text{for all }t\in[0,b]. Proof. Let M_{n}(t)=\int_{0}^{n}N_{\omega}d\omega and G_{n}(t)=\int_{0}^{n}g_{\omega}d\omega . By integrating ( 5 ) with respect to \omega , we have M_{n}(t)=e^{\psi(t)}\int_{0}^{n}N_{\omega}(0)d\omega+e^{\psi(t)}\int_{0}^{t}e^{-\psi(t)}\int_{0}^{n}\int_{0}^{\infty}N_{\omega+\mu}N_{\mu}d\mu d\omega dt+e^{\psi(t)}\int_{0}^{t}e^{-\psi(t)}\int_{0}^{n}g_{\omega}(t)d\omega dt. Now, as