Zusatz: Übersicht über die behandelten DGLs

Alle Funktionen hängen von \((x, t) \in \Omega _T\) ab, soweit nicht anders erklärt. Es ist \(\div := \div _x\) und \(\Delta := \Delta _x\).

DGL

Name

Herleitung

\(\partial _t u + \div F = G\)

Transport-Reaktionsgleichung

Massenbilanz in Kontrollvolumen, \(F(x,t) \in \real ^d\) Fluss, \(G(x,t) \in \real \) Konz.gewinn

\(\partial _t u = G(x, t, u(x,t))\)

parametrisierte ODE

aus TRGL: \(F :\equiv 0\), \(G\) \(u\)-abhängig

\(\partial _t u + \div (vu) = 0\)

Advektionsgleichung

aus TRGL: \(F := vu\) mit \(v \in \C ^1(\Omega _T, \real ^d)\) Geschw.feld, \(G :\equiv 0\)

\(\partial _t u + \div F(u) = 0\)

nicht-lineare Konvektionsgl.

aus TRGL: \(F := F(u)\), \(G :\equiv 0\)

\(\partial _t u + \partial _x (v(u) \cdot u) = 0\)

Konvektionsgleichung

aus nicht-linearer Konvektionsgl.: \(d := 1\), \(F(u) := v(u) \cdot u\)

\(\partial _t u + \partial _x(\frac {1}{2} u^2) = 0\)

Burgersgleichung

aus nicht-linearer Konvektionsgl.: \(d := 1\), \(F(u) := \frac {1}{2} u^2\)

\(\partial _t u - \div (a(x) \nabla u) = 0\)

allg. Diffusionsgleichung

aus TRGL: \(F := -a(x) \nabla u\) (Ficksches Gesetz) mit \(a \in \C ^1(\Omega )\) Diff.koeff., \(G :\equiv 0\)

\(\partial _t u - \Delta u = 0\)

Diffusionsgleichung/instat. WLG

aus allg. Diffusionsgleichung: \(a(x) :\equiv 1\)

\(-\Delta u = 0\)

Laplace-Gleichung

aus instat. Wärmeleitungsgleichung mit \(t \to \infty \) und \(u(\cdot , t) \to \overline {u}(\cdot ) \in \C ^2(\overline {\Omega })\) glm.

\(-\Delta u = f\)

Poisson-Gleichung

aus Laplace-Gleichung mit \(G := f(x)\)

\(-\div _x(\nabla _p L(\nabla u, u, x)) + \partial _z L(\nabla u, u, x) = 0\)

Euler-Lagrange-Gleichung

PDE für Lösung \(u\) des Variationsproblems \(I(u) \le I(w) := \int _\Omega L(\nabla w, w, x)\dx \)

\(-a_{11}(x) u’’(x) + c(x) u(x) = f(x)\)

Sturm-Liouville-Problem

aus ELGL: \(L(p, z, x) := \frac {1}{2} p^\tp A(x) p + \frac {1}{2} c(x) z^2 - z f(x)\), \(d := 1\), \(a_{11}(x) > 0\), \(c(x) > 0\)

\(\partial _t^2 u - c^2 \Delta u = 0\)

Wellengleichung

aus ELGL: \(L(p, z, x) := \frac {c^2}{2} \sum _{i=1}^d |p_i|^2 - \frac {1}{2} |p_{d+1}|^2\) aus Hamilton-Prinzip

\(-\Delta u + W’(u) = 0\)

stat. Allen-Cahn-Gleichung

aus ELGL: \(L(p, z, x) := W(z) + \frac {1}{2} \norm {p}^2\) mit z. B. \(W(z) := (z^2 - 1)^2\)

Zusatz: Übersicht über die Aussagen über PDE-Klassen

PDE

Problem

Definition/Satz

Voraussetzungen/Aussage

Advektions-gleichung konstante Adv.geschw.

Definition

\(\Omega := \real ^d\), \(T := \infty \), \(b \in \real ^d\), \(u_0 \in \C ^1(\Omega )\) \(\implies \) \(\partial _t u + \div (bu) = 0\) in \(\Omega _T\), \(u(\cdot , 0) = u_0\) in \(\Omega \)

Translationsinv.

\(\forall _{(x, t) \in \Omega _T} \forall _{s \in (-t, T-t)}\; \frac {\d }{\ds } u(x + bs, t + s) = 0\)

Ex. + Eind.

\(u(x, t) := u_0(x - bt)\) eind. kl. Lsg.

\(L^\infty \)-Stabilität

\(u_0 \in \C ^1(\Omega ) \cap L^\infty (\Omega ) \implies \forall _{t \in (0, T)}\; \norm {u(\cdot , t)}_{L^\infty } \le \norm {u_0}_{L^\infty }\)

Max.-/Min.prinzip

\(u_0 \in \C ^1(\Omega ) \cap L^\infty (\Omega ) \implies \forall _{(x, t) \in \Omega _T}\; \inf _{\overline {x} \in \Omega } u_0(\overline {x}) \le u(x, t) \le \sup _{\overline {x} \in \Omega } u_0(\overline {x})\)

st. Abh. von \(u_0\)

\(u_0, u_0’ \in \C ^1(\Omega ) \cap L^\infty (\Omega ) \implies \forall _{t \in (0, T)}\; \norm {u(\cdot , t) - u’(\cdot , t)}_{L^\infty } \le \norm {u_0 - u_0’}_{L^\infty }\)

keine st. Abh. von \(b\)

\(\lnot [\forall _{t \in (0, T)} \exists _{C(t) > 0} \forall _{u_0 \in \C ^1(\Omega ) \cap L^\infty (\Omega ), \norm {u_0}_{L^\infty } \le 1} \forall _{b, b’ \in \real }\; \norm {u(\cdot , t) - u’(\cdot , t)}_{L^\infty } \le C(t) \norm {b - b’}]\)

Reaktions-/ Quellterm

Definition

\(q \in \C ^0(\Omega _T)\) \(\implies \) \(\partial _t u + \div (bu) = q\) in \(\Omega _T\), \(u(\cdot , 0) = u_0\) in \(\Omega \)

Ex. + Eind.

\(u(x, t) := u_0(x - bt) + \int _0^t q(x + (s-t)b, s)\ds \) eind. kl. Lsg.

nicht-lineare Konvektion

Definition

\(\Omega := \real \), \(T > 0\), \(f \in \C ^2(\real )\), \(u_0 \in \C ^1(\Omega )\) \(\implies \) \(\partial _t u + \partial _x(f(u)) = 0\) in \(\Omega _T\), \(u(\cdot , 0) = u_0\) in \(\Omega \)

lokale Ex.

\(\norm {f’’}_\infty , \norm {u_0’}_\infty < \infty \implies \forall _{\overline {x} \in \real } \exists _{\varepsilon > 0} \exists _{T > 0} \exists _{u \in \C ^1(B_\varepsilon (\overline {x}) \times (0, T))}\; [\text {$u$ kl. Lsg.}]\), \(u(x, t) = u_0(x - tf’(u(x, t)))\)

Poisson-Gleichung Laplace-Gleichung

Definition

\(\Omega \subset \real ^d\) \(\implies \) \(-\Delta u = 0\) in \(\Omega \)

MW-Eigenschaft

\(u \in C^2(\Omega )\) harm., \(x \in \Omega \), \(r > 0\), \(\overline {B_r(x)} \subset \Omega \) \(\implies \fint _{B_r(y)} u(y) \dy = u(x) = \fint _{\partial B_r(x)} u(y) \dsigma (y)\)

Max.prinzip

\(\Omega \) offen, beschr., \(u \in \C ^2(\overline {\Omega })\) harm. \(\implies \max _{x \in \overline {\Omega }} u(x) = \max _{x \in \partial \Omega } u(x)\)

verallg. Max.prinz.

\(\Omega \) offen, beschr., \(u \in \C ^2(\Omega ) \cap \C ^0(\overline {\Omega })\), \(-\Delta u = f \le 0\) \(\implies \) \(u\) nimmt Max. auf dem Rand an

Vgl.prinzip

\(\Omega \) offen, beschr., \(u, v \in \C ^2(\Omega ) \cap \C ^0(\overline {\Omega })\), \(-\Delta u \le -\Delta v\) in \(\Omega \), \(u \le v\) auf \(\partial \Omega \) \(\implies \) \(u \le v\) in \(\Omega \)

Regularität

\(\Omega := \real ^d\), \(u \in \C ^2(\Omega )\) harm. \(\implies u \in \C ^\infty (\Omega )\)

Fundamentallsg.

\(\Omega := \real ^d \setminus \{0\}\), \(d > 1\) \(\implies \Phi \in \C ^\infty (\Omega )\), \(\Phi (x) := -\frac {1}{2\pi } \cdot \ln (\norm {x})\) für \(d = 2\), \(\Phi (x) := \frac {1}{(d-2)\omega _d} \cdot \frac {1}{\norm {x}^{d-2}}\) für \(d \ge 3\)

Eigenschaften

\(\int _{B_\varepsilon (0)} \Phi (x)\dx \to 0\), \(\Phi \in L^1_\loc (\real ^d)\), \(\Phi (\varepsilon e_1) \varepsilon ^{d-1} \to 0\), \(\forall _{\varepsilon >0}\; \int _{\partial B_\varepsilon (0)} \nabla \Phi (x) \cdot n\dsigma (x) = -1\)

Poisson-Gleichung

Definition

\(\Omega \subset \real ^d\), \(f\colon \Omega \to \real \) \(\implies \) \(-\Delta u = f\) in \(\Omega \)

Rotationsinv.

\(u \in \C ^2(\Omega )\) kl. Lsg., \(O \in \real ^{d \times d}\) orth., \(\Omega = O\Omega \), \(f = f \circ O\) \(\implies v \in \C ^2(\Omega )\) kl. Lsg., \(v(x) := u(Ox)\)

Faltungslösung

\(\Omega := \real ^d\), \(d \ge 2\), \(f \in \C ^2_0(\Omega )\) \(\implies u := \Phi \ast f\) kl. Lsg.

Poisson-RWP

Definition

\(\Omega \subset \real ^d\) offen, beschr., \(f \in \C ^0(\Omega )\), \(g \in \C ^0(\partial \Omega )\) \(\implies \) \(-\Delta u = f\) in \(\Omega \), \(u = g\) auf \(\partial \Omega \)

Eind.

es gibt höchstens eine kl. Lsg. \(u \in \C ^2(\Omega ) \cap \C ^0(\overline {\Omega })\)

st. Abh. von \(g\)

\(g, g’ \in \C ^0(\partial \Omega ) \implies \norm {u - u’}_\infty \le \norm {g - g’}_\infty \)

st. Abh. von \(f\)

\(f, f’ \in C^0(\Omega ) \implies \norm {u - u’}_\infty \le C \norm {f - f’}_\infty \), \(C := \frac {R^2}{2}\), \(R := \sup _{x \in \Omega } \norm {x}\)

PDE

Problem

Definition/Satz

Voraussetzungen/Aussage

Diffusions-gleichung AWP

Definition

\(\Omega \subset \real ^d\), \(T > 0\), \(u_0\colon \Omega \to \real \) \(\implies \) \(\partial _t u - \Delta u = 0\) in \(\Omega _T\), \(u(\cdot , 0) = u_0\) in \(\Omega \)

Skal.inv.

\(\Omega := \real ^d\), \(T := \infty \), \(u \in \C ^2(\Omega _T)\) kl. Lsg., \(\lambda \in \real \) \(\implies u_\lambda \) kl. Lsg., \(u_\lambda (x, t) := u(\lambda x, \lambda ^2 t)\)

Fundamentallsg.

\(\Omega := \real ^d\), \(T := \infty \) \(\implies \Phi \in \C ^\infty (\Omega _T)\), \(\Phi (x, t) := \frac {1}{(4\pi t)^{d/2}} e^{-\norm {x}^2/(4t)}\)

Faltungslösung

\(\Omega := \real ^d\), \(T := \infty \), \(u_0 \in L^\infty (\Omega )\) \(\implies u \in \C ^\infty (\Omega _T)\), \(u(\cdot , t) := \Phi (\cdot , t) \ast u_0\) kl. Lsg.,
für \(u_0 \in \C ^0(\real )\) gilt \(\forall _{\overline {x} \in \Omega }\; \lim _{(x, t) \to (\overline {x}, 0)} u(x, t) = u_0(\overline {x})\), \(\forall _{t > 0}\; \norm {u(\cdot , t)}_{L^\infty } \le \norm {u_0}_{L^\infty }\)

ARWP

Definition

\(g\colon \partial \Omega \times (0, T) \to \real \) \(\implies \) \(\partial _t u - \Delta u = 0\) in \(\Omega _T\), \(u(\cdot , 0) = u_0\) in \(\Omega \), \(u = g\) auf \(\partial \Omega \times (0, T)\)

Max.prinzip

\(u\) nimmt Maximum auf parabolischem Rand \(\Gamma := (\Omega \times \{0\}) \cup (\partial \Omega \times [0, T])\) an

inhom. ARWP

Definition

\(f\colon \Omega _T \to \real \) \(\implies \) \(\partial _t u - \Delta u = f\) in \(\Omega _T\), \(u(\cdot , 0) = u_0\) in \(\Omega \), \(u = g\) auf \(\partial \Omega \times (0, T)\)

Eind.

\(\Omega \subset \real ^d\) Lipschitz \(\implies \) es gibt höchstens eine kl. Lsg.

Konv. gg. stat. Lsg.

\(\Omega \subset \real ^d\) Lipschitz, \(f, g\) zeitunabh., \(-\Delta \overline {u} = f\) in \(\Omega \), \(\overline {u} = g\) auf \(\partial \Omega \) \(\implies \norm {u(\cdot , t) - \overline {u}}_{L^2} \le e^{-t/c_p} \norm {u_0 - \overline {u}}_{L^2}\)

Wellen-gleichung AWP

Definition

\(\Omega := \real ^d\), \(T > 0\), \(c > 0\), \(u_0 \in \C ^2(\Omega )\), \(v_0 \in \C ^1(\Omega )\) \(\implies \) \(\partial _t^2 u - c^2 \Delta u = 0\) in \(\Omega _T\), \(u(\cdot , 0) = u_0\), \(\partial _t u(\cdot , 0) = v_0\) in \(\Omega \)

Ex. + Eind.

\(d := 1 \implies u(x, t) := \frac {1}{2} (u_0(x+ct) + u_0(x-ct)) + \frac {1}{2c} \int _{x-ct}^{x+ct} v_0(s)\ds \) eind. kl. Lsg.

\(L^\infty \)-Stabilität

\(d := 1\), \(u_0 \in \C ^2(\Omega ) \cap L^\infty (\Omega )\), \(v_0 \in \C ^1(\Omega ) \cap L^1(\Omega )\) \(\implies \forall _{t \ge 0}\; \norm {u(\cdot , t)}_{L^\infty } \le \norm {u_0}_{L^\infty } + \frac {1}{2c} \norm {v_0}_{L^1}\)

st. Abh. von \(u_0, v_0\)

\(d := 1\), \(u_0, \overline {u_0} \in \C ^2 \cap L^\infty \), \(v_0, \overline {v_0} \in \C ^1 \cap L^1\) \(\implies \forall _{t \ge 0}\; \norm {u(\cdot , t) - \overline {u}(\cdot , t)}_{L^\infty } \le C \left (\norm {u_0 - \overline {u_0}}_{L^\infty } + \norm {v_0 - \overline {v_0}}_{L^1}\right )\)

Abh.kegel

\(d := 1\), \((x_0, t_0) \in \Omega _T\), \(\forall _{|x - x_0| \le ct_0}\; u_0(x) = v_0(x) = 0\) \(\implies \) \(u(x,t) = 0\) für \(t \in [0, t_0]\), \(|x - x_0| \le c(t_0 - t)\)

inhom. ARWP

Definition

\(\Omega \subset \real ^d\), \(f\colon \Omega _T \to \real \), \(g\colon \partial \Omega \times (0, T) \to \real \), \(u_0, v_0\colon \Omega \to \real \)
\(\implies \) \(\partial _t^2 u - c^2 \Delta u = f\) in \(\Omega _T\), \(u(\cdot , 0) = u_0\) in \(\Omega \), \(\partial _t u(\cdot , 0) = v_0\) in \(\Omega \), \(u = g\) auf \(\partial \Omega \times (0, T)\)

Eind.

\(\Omega \subset \real ^d\) Lipschitz \(\implies \) es gibt höchstens eine kl. Lsg.