Partial Differential Equations
By Shaharuddin Salleh, B. Zaltzman, Albert Y. Zomaya, S. A. Bakar. 2018
3,416 citations recorded by the source corpus.
In the following exercises we assume U is open, bounded set, with smooth boundary, and T > 0. Exercise 16 still has some gap to be overcame. The difficult exercise 9 is solved by mimicking a proof in a paper of Brezis-Evans on 2016/07/31. 1. Prove there is at most one smooth solution of this initial/boundary-value problem for the heat equation with Neumann boundary conditions ut −∆u = f in U × (0, T ) ∂u ∂ν = 0 on ∂U × [0, T ] u = g on U × {t = 0}. (1) Proof. Consider E(t) = ∫ U (v(x, t)) dx, where v solves the heat equation with all corresponding boundary conditions are zero. Then we see E ′(t) ≤ 0 and E(0) = 0 ≤ E(t), ∀t. So E(t) ≡ 0 and hence v ≡ 0. 1 2 Prove the following weak maximum principle for (1) with f = 0: If g ≥ 0 in U , then u ≥ 0 in UT . Proof. The proof strategy is the same as the Dirichlet problem except the contradiction argument for ∂U needs the Hopf boundary lemma. ∗Department of Math., National Taiwan University. Email: d04221001@ntu.edu.tw 1 Note that we may assume g ≥ c > 0 by considering u+ c and letting c→ 0 in the final step. Consider u ,δ(x, t) = u(x, t)+ t+δ[|x|+h(x)], where , δ > 0 and h(x) is the harmonic function which equals to −|x| on ∂U (whose existence can be proved by Perron’s method or using the Riesz representation theorem with elliptic regularity theory as Chapter 6). Since |x| + h(x) is bounded (due to its continuity on U), u ,δ(x, 0) ≥ c − δ‖|x| + h(x)‖∞ ≥ 0 for all small δ > 0 (independent of ). Suppose inf(x,t)∈UT u ,δ(x, t) < 0. By continuity of u ,δ, we know the infimum is attained at (x0, t0). It’s obvious that t0 > 0. If (x0, t0) ∈ U × (0, T ], then 0 ≥ (∂t −∆)u ,δ(x0, t0) = − 2nδ > 0 for all small δ > 0. So (x0, t0) ∈ ∂U × (0, T ]. However, the minimality at (x0, t0) implies ∂u ,δ ∂ν (x0, t0) ≤ 0. However, the Hopf boundary lemma implies ∂ ∂ν (|x| 2 + h(x))(x0) > 0 and hence ∂u ,δ ∂ν (x0, t0) > 0. Therefore inf(x,t)∈UT u ,δ(x, t) ≥ 0 and hence inf(x,t)∈UT u(x, t) ≥ 0 by letting , δ → 0. 2. Proof. Since d dt 1 2 ‖u(·, t)‖2L2 = − ∫ U u(−∆u) ≤ −λ1‖u‖2L2 , basic Gronwall’s inequality shows the desired inequality. 3. Proof. 0 = ∫ UT (ut + Lu)v d(x, t) = ∫ U ∫ T 0 utv dt dx+ ∫ U ∫ T 0 (Lu)v dt dx = ∫ U u(x, T )v(x, T )− u(x, 0)v(x, 0) dx− ∫ UT u(vt − L∗u) d(x, t) 4. Proof. Let {wk} be the family of all eigenfunctions of −∆ with zero Dirichlet boundary conditions. It’s an orthogonal basis of H 0 (U) (See Evans, step 3 in page 357. By Poincaré’s inequality, β‖um‖2H1 ≤ ‖Dum‖2L2 = m ∑ k=1 dm ∫
Subfields: Computational Theory and Mathematics, Numerical Analysis
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computational Theory and Mathematics
- Youshan Tao
- Miroslav Krstic
- Rafael Vazquez
- Florent Di Meglio
- M. Mitrea
- Giuseppe Viglialoro
- Michael Winkler
- Long Hu
In Numerical Analysis
- Miroslav Krstic
- Rafael Vazquez
- Sifan Wang
- Gordan vZitkovi'c
- Mohammad F. Al-Jamal
- Eduardo Casas
- Gustavo Ponce
- Adam S. Sikora
Citations in the corpus
- Physics informed deep learning for computational elastodynamics without labeled data (2020, background): “This owes to the nature of hyperbolic PDEs as further explained in Remark 1, which has been elaborated in [59].”
- Understanding and mitigating gradient pathologies in physics-informed neural networks (2020, background): “As we know, in lack of proper restrictions such as boundary conditions or initial conditions, a PDE system may have infinitely many solutions [29].”
- Kinetic formulation and global existence for the Hall-Magneto-hydrodynamics system (2011, neutral): “By Gagliardo-Nirenberg’s inequality [10],”
- Conjugate gradient method for the Robin inverse problem associated with the Laplace equation (2007, background): “Therefore, the Holmgren’s unique continuation theorem [21] concludes u(x)≡ 0, x∈ which is in contradiction with the assumption that h(x) is not identically zero.”
- Schauder Estimates for Elliptic and Parabolic Equations* (2006, neutral): “5) can be found in [E2].”
- On the lack of null-controllability of the heat equation on the half-line (2000, neutral): “1) has multiple smooth solutions (see [13], pp.”
- Nonlinear normal modes and spectral submanifolds: existence, uniqueness and use in model reduction (2016, background): “This argument just reproduces the classic local existence and uniqueness result for linear first-order PDEs (see, e.g., Evans [16]).”
- The Spotting Distribution of Wildfires (2016, methodology): “Our use of the spotting distribution is similar to the Green’s-function approach that is used in the study of linear partial differential equations [50].”
- A Nonlinear Case of the 1-D Backward Heat Problem: Regularization and Error Estimate (2007, neutral): “n=1 φn sinnx, where we recall un(0) = 2 π 〈u(x, 0), sinnx〉 (see [5]).”
- Poincaré and Brunn-Minkowski inequalities on the boundary of weighted Riemannian manifolds (2017, neutral): “[73, 58]); for more general situations (weaker regularity of”