The Robin heat kernel and its expansion via Robin eigenfunctions
Yifeng Meng, Kui Wang · 2025
Read The Robin heat kernel and its expansion via Robin eigenfunctions by Yifeng Meng, Kui Wang (2025) 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-2505-15092.txt.
Introduction
Yifeng Meng and Kui Wang build the heat kernel for a Robin edge rule on curved spaces with a rim, in a 2025 paper. The kernel tells how heat spreads when the edge rule blends the value at the rim with its outward slope. Classic work had settled the fixed-value and no-flow edge rules.
The pair prove the kernel exists, is the only one, and unfolds as a sum over the rule's own tones and shapes. For kind parameter values they argue straight from trace bounds and shape estimates. Cruel values below zero need fresh tricks to cap the shapes' peaks. The authors say the gain is a full spectral picture for the less-studied cruel regime.
Catalog record
| CREATOR | Yifeng Meng, Kui Wang |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 27 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-15092
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
The Robin heat kernel and its expansion via Robin eigenfunctions Yifeng Meng, Kui Wang Abstract . We prove the existence and uniqueness of the Robin heat kernel on compact Riemannian manifolds with smooth boundary for Robin parameter α ∈ ℝ \alpha\in\mathbb{R} , expressed as a spectral expansion in terms of Robin eigenvalues and eigenfunctions. For the non-negative parameter regime ( α ≥ 0 \alpha\geq 0 ), we present a direct proof based on trace Sobolev inequalities and eigenfunction estimates. The case of negative parameters ( α < 0 \alpha<0 ) requires novel analytical techniques to handle L ∞ L^{\infty} estimates of Robin eigenfunctions, addressing challenges not present in the non-negative case. Our result extends the the classical Dirichlet and Neumann cases to the less-studied negative parameter regime. 2020 Mathematics Subject Classification 35P15, 53C21 1. Introduction Let M^{m} be an m -dimensional compact Riemannian manifold with smooth boundary, and we consider the heat equation (1.1) \displaystyle u_{t}(x,t)-\Delta u(x,t)=0,\qquad(x,t)\in M\times(0,T), with the Robin boundary condition (1.2) \displaystyle\frac{\partial}{\partial\nu}u+\alpha u=0,\qquad(x,t)\in\partial M\times(0,T), where \nu denotes the unit outward normal to \partial M , and \alpha\in R is the Robin parameter. It is well known that for the Neumann boundary ( \alpha=0 ) and the Dirichlet boundary ( \alpha=+\infty ), the solution of ( 1.1 ) with initial data u(x,0)=u_{0}(x) can be expressed as u(x,t)=\int_{M}H_{\alpha}(x,y,t)u_{0}(y)dy. where H_{0}(x,y,t) and H_{+\infty}(x,y,t) represent Neumann and Dirichlet heat kernels, respectively (see [ 1 ] , [ 12 , Chapter 10] ). For further discussion on some space-time boundary conditions, we refer to [ 11 ] . Heat kernels play an important role in the study of partial differential equations and geometric analysis, as evidenced by [ 2 , 15 ] and the comprehensive treatments [ 5 , 9 , 18 ] . Recent progress on the heat kernel theory can be found in [ 3 , 4 , 13 , 14 , 17 ] . To state our main result, we first introduce the Robin eigenvalue problem: let \lambda_{i,\alpha} and \phi_{i,\alpha}(x) denote the Robin eigenvalues and eigenfunctions of Laplacian on M with parameter \alpha\in\mathbb{R} , defined by (1.3) \displaystyle\begin{cases}-\Delta\phi_{i,\alpha}=\lambda_{i,\alpha}\phi_{i,\alpha},&\quad x\in M,\\ \frac{\partial}{\partial\nu}\phi_{i,\alpha}+\alpha\phi_{i,\alpha}=0,&\quad x\in\partial M,\end{cases} where the eigenvalues satisfy \lambda_{1,\alpha}<\lambda_{2,\alpha}\leq\lambda_{3,\alpha}\leq\cdots\to+\infty, For details, see [ 10 , Chapter 4] . Denote by \Delta_{\alpha} be the Laplacian with Robin boundary condition ( 1.2 ). When \alpha\geq 0 , the operator \Delta_{\alpha} is non-negative, then Theorem 2.1.4 of [ 7 ] implies the Robin heat kernel has the following expansion formula H_{\alpha}(x,y,t)=\sum_{i=1}^{\infty}e^{-\lambda_{i,\alpha}t}\phi_{i,\alpha}(x)\phi_{i,\alpha}(y). The expansion formula ( 1.4 ) is well-established for Dirichlet and Neumann heat kernels, as well as for heat kernels on closed manifolds, see [ 1 , 7 , 12 ] . However, the case \alpha<0 has been less studied. In this paper, we address this remaining case. Theorem 1.1 . Let M be a compact Riemannian manifold with smooth boundary, \alpha\in\mathbb{R} , \lambda_{i,\alpha} be the Robin eigenvaules defined by ( 1.3 ), and \phi_{i,\alpha} be the corresponding normalized eigenfunctions. Denote with (1.4) \displaystyle H_{\alpha}(x,y,t)=\sum_{i=1}^{\infty}e^{-\lambda_{i,\alpha}t}\phi_{i,\alpha}(x)\phi_{i,\alpha}(y). Then, H_{\alpha}(x,y,t) is well defined on M\times M\times(0,\infty) , and is the unique kernel such that: for any u_{0}(x)\in L^{2}(M) , the solution of equation ( 1.1 ) with the Robin boundary condition and initial condition u(x,0)=u_{0} is given by \displaystyle u(x,t)=\int_{M}H_{\alpha}(x,y,t)u_{0}(y)dy. The proof of our main theorem combines techniques from spectral theory, elliptic regularity, and geometric analysis. For \alpha>0 , we provide a direct proof using a trace Sobolev inequality and iteration methods. The case \alpha<0 presents additional challenges due to the negativity of the principal eigenvalue; we overcome this through a careful decomposition of the first eigenfunction. The uniqueness follows from the maximum principle for the Robin heat equation (Theorem 2.1 ), which extends classical results for Dirichlet and Neumann boundary conditions. This paper is organized as follows: Section 2 presents preliminary results on Robin eigenvalues, maximum principles, and trace inequalities. Section 3 contains the proof of our main theorem, with separate treatments for positive and negative Robin parameters. 2. Preliminaries This section establishes the foundational results necessary for our analysis of the Robin heat kernel. We adopt the following notation throughout: for any function f\in L^{q}(M) , its L^{q} norm is denoted by \|f\|_{q}:=(\int_{M}|f(x)|^{q}\,dx)^{1/q}. 2.1. Robin Eigenvalue Problem. Let M be a compact m -dimensional Riemannian manifold with smooth boundary, \lambda_{i,\alpha} ( i=1,2,\cdots ) be the Robin eigenvalues, and \phi_{i,\alpha} be normalized eigenfunctions such that \|\phi_{i,\alpha}\|_{2}=1 . It is well known that the eigenvalue problem ( 1.3 ) is equivalent to the following variation problem (2.1) {}\lambda_{i,\alpha}(M)=\inf_{\begin{subarray}{c}H\subset H^{1}(M),\\ \dim H=i\end{subarray}}\sup_{0\neq u\in H}\frac{\int_{M}|\nabla u|^{2}dx+\alpha\int_{\partial M}u^{2}dS}{\int_{M}u^{2}dx}, and particularly (2.2) {}\lambda_{1,\alpha}(M)=\inf_{0\neq u\in H^{1}(M)}\frac{\int_{M}|\nabla u|^{2}\,dx+\alpha\int_{\partial M}u^{2}\,dS}{\int_{M}u^{2}\,dx}, where dS is the induced measure on \partial M . Moreover, the eigenfunctions \phi_{i,\alpha} ( i=1,2,\cdots ) form a complete orthonormal basis for L^{2}(M) . If M is connected, the Krein-Rutman Theorem guarantees the simplicity of \lambda_{1,\alpha} and strict positivity of its eigenfunction (see [ 10 , Section 4.2] ). When \alpha<0 , we obtain the following additional information. Proposition 2.1 . Suppose \alpha<0 , and denote by \phi_{1,\alpha} be the positive normalized eigenfunction with respect to \lambda_{1,\alpha} . Then, \lambda_{1,\alpha}(M)<0 and \inf_{M}\phi_{1,\alpha}>0 . Proof. Taking u=1 as a trial function in ( 2.2 ) yields \lambda_{1,\alpha}(M)\leq\frac{\alpha\int_{\partial M}\,dS}{\int_{M}\,dx}<0. Suppose \inf_{M}\phi_{1,\alpha}=0 , then there exists a x_{0}\in\partial M such that \phi_{1,\alpha}(x_{0})=0 . The Robin boundary condition ( 1.2 ) gives \partial_{\nu}\phi_{1,\alpha}(x_{0})=0 , which contradicts with the Hopf lemma. Hence \inf_{M}\phi_{1,\alpha}>0 . \square 2.2. Maximum Principles and Uniqueness The Robin heat equation satisfies the following maximum principle: Theorem 2.1 . Let M be a compact manifold with smooth boundary. Suppose that u(x,t)\in C^{2,1}(M\times[0,\infty)) satisfies (2.3) \displaystyle\begin{cases}\left(\partial_{t}-\Delta\right)u\geq 0,&\quad(x,t)\in M\times(0,\infty),\\ \frac{\partial}{\partial\nu}u+\alpha u\geq 0,&\quad(x,t)\in\partial M\times(0,+\infty),\\ u(x,0)\geq 0,&\quad x\in M.\end{cases} Then, u(x,t)\geq 0 for all x\in M and t>0 . Proof. Let \beta=\min\{\alpha,0\}-1 , and denote by \lambda_{1,\beta} and \phi_{1,\beta} the first Robin eigenvalue and its corresponding positive normalized eigenfunction. Clearly w(x,t):=e^{-\lambda_{1,\beta}t}\phi_{1,\beta}(x) is strictly positive and satisfies (2.4) \displaystyle\begin{cases}\left(\partial_{t}-\Delta\right)w=0,&\quad(x,t)\in M\times(0,\infty),\\ \frac{\partial}{\partial\nu}w+\beta w=0,&\quad x\in\partial M.\\ \end{cases} For any \varepsilon>0 , define v_{\varepsilon}(t,x):=u(t,x)+\varepsilon w(t,x) . By ( 2.3 ) and ( 2.4 ), we derive that \displaystyle\begin{cases}\left(\partial_{t}-\Delta\right)v_{\varepsilon}\geq 0,&\quad(x,t)\in M\times(0,\infty),\\ \frac{\partial}{\partial\nu}v_{\varepsilon}+\alpha v_{\varepsilon}\geq\varepsilon(\alpha-\beta)w>0,&\quad x\in\partial M,\\ v_{\varepsilon}>0,&\quad t=0.\end{cases} We claim that v_{\varepsilon}(x,t)>0 for all ((x,t)\in[0,\infty)\times M . If not, let t_{0} be the first time such that v_{\varepsilon}(x_{0},t_{0})=0 for some x_{0}\in M\cup\partial M , and by strong parabolic maximal principle, we have x_{0}\in\partial M . Then we have \displaystyle 0\geq\partial_{\nu}v_{\varepsilon}(x_{0},t_{0})=-\alpha v_{\varepsilon}(x_{0},t_{0})+\varepsilon(\alpha-\beta)w(x_{0},t_{0})=\varepsilon(\alpha-\beta)w(x_{0},t_{0}), contradicting with w(x_{0},t_{0})>0 . Hence u(x,t)+\varepsilon w(x,t)>0,\quad(x,t)\in M\times(0,\infty) for all \varepsilon>0 . Then letting \varepsilon\to 0 , we have u(x,t)\geq 0, proving the theorem. \square The following corollary comes true directly from Theorem 2.1 . Corollary 2.2 (Uniqueness and Positivity) . The solution to the Robin heat equation with given initial data is unique, and the kernel function H_{\alpha}(x,y,t)\geq 0 almost everywhere if exists. 2.3. Trace Sobolev Inequalities The analysis of Robin boundary conditions requires careful control of boundary terms. We establish the following fundamental inequalities: Lemma 2.1 . Let M be a compact manifold with smooth boundary. Then there exists a constant C_{1} , depending on M , such that for all u\in H^{1}(M) it holds (2.5) \displaystyle\int_{\partial M}u^{2}dS\leq C_{1}\left(\|\nabla u\|_{2}\|u\|_{2}+\|u\|_{2}^{2}\right). Proof. Recall from standard trace theorem [ 8 , Chapter 5] that for all v\in W^{1,1}(M) it holds \int_{\partial M}|v|dS\leq C(M)\left(\int_{M}|\nabla v|+\int_{M}|v|dx\right), and taking v=u^{2} gives ( 2.5 ). \square