Difference between revisions of "Schauder method"
Ulf Rehmann (talk | contribs) m (tex encoded by computer) |
m (fix tex) |
||
Line 18: | Line 18: | ||
Lu \equiv \sum _ {i, j= 1 } ^ { n } | Lu \equiv \sum _ {i, j= 1 } ^ { n } | ||
a ^ {ij} ( x) u _ {x _ {i} x _ {j} } + | a ^ {ij} ( x) u _ {x _ {i} x _ {j} } + | ||
− | \sum _ { j= } | + | \sum _ { j=1 } ^ { n } b ^ {j} ( x) u _ {x _ {j} } + b |
( x) u = f( x), | ( x) u = f( x), | ||
$$ | $$ | ||
Line 43: | Line 43: | ||
$$ | $$ | ||
\| u \| _ {C _ {1+ \alpha } ( \Omega ) } = \| u \| _ {C _ \alpha ( | \| u \| _ {C _ {1+ \alpha } ( \Omega ) } = \| u \| _ {C _ \alpha ( | ||
− | \Omega ) } + \sum _ { i= } | + | \Omega ) } + \sum _ { i=1 } ^ { n } \| u _ {x _ {i} } \| _ {C _ \alpha ( \Omega ) } , |
$$ | $$ | ||
$$ | $$ | ||
\| u \| _ {C _ {2+ \alpha } ( \Omega ) } = \| u \| _ {C _ {1+ \alpha } | \| u \| _ {C _ {2+ \alpha } ( \Omega ) } = \| u \| _ {C _ {1+ \alpha } | ||
− | ( \Omega ) } + \sum _ { i,j= } | + | ( \Omega ) } + \sum _ { i,j=1 } ^ { n } \| u _ {x _ {i} x _ {j} } \| _ {C _ \alpha ( \Omega ) } . |
$$ | $$ | ||
Line 88: | Line 88: | ||
$$ | $$ | ||
− | for the Laplace operator $ \Delta = \sum _ {i=} | + | for the Laplace operator $ \Delta = \sum _ {i=1} ^ {n} \partial ^ {2} / \partial x _ {i} ^ {2} $. |
5) Without loss of generality one may assume that , | 5) Without loss of generality one may assume that \phi ( x) \equiv 0 , |
Latest revision as of 16:50, 4 January 2021
A method for solving boundary value problems for linear uniformly-elliptic equations of the second order, based on a priori estimates and the continuation method (see also Continuation method (to a parametrized family)).
Schauder's method of finding a solution to the Dirichlet problem for a linear uniformly-elliptic equation
\tag{1 } Lu \equiv \sum _ {i, j= 1 } ^ { n } a ^ {ij} ( x) u _ {x _ {i} x _ {j} } + \sum _ { j=1 } ^ { n } b ^ {j} ( x) u _ {x _ {j} } + b ( x) u = f( x),
given in a bounded domain \Omega of a Euclidean space of points x= ( x _ {1} \dots x _ {n} ) and with a coefficient b( x) \leq 0 , can be described in the following way.
1) The spaces C _ \alpha ( \Omega ) , C _ {1+ \alpha } ( \Omega ) and C _ {2+ \alpha } ( \Omega ) are introduced as sets of functions u = u( x) with finite norms
\| u \| _ {C _ \alpha ( \Omega ) } = \sup _ {x \in \Omega } | u( x) | + \sup _ {x,y } \frac{u( x)- u( y) }{| x- y | ^ \alpha } ,\ \ 0 < \alpha < 1,
\| u \| _ {C _ {1+ \alpha } ( \Omega ) } = \| u \| _ {C _ \alpha ( \Omega ) } + \sum _ { i=1 } ^ { n } \| u _ {x _ {i} } \| _ {C _ \alpha ( \Omega ) } ,
\| u \| _ {C _ {2+ \alpha } ( \Omega ) } = \| u \| _ {C _ {1+ \alpha } ( \Omega ) } + \sum _ { i,j=1 } ^ { n } \| u _ {x _ {i} x _ {j} } \| _ {C _ \alpha ( \Omega ) } .
2) It is assumed that the boundary \sigma of the domain \Omega is of class C _ {2 + \alpha } , i.e. each element \sigma _ {x} of the ( n- 1) - dimensional surface \sigma can be mapped on a part of the plane by a coordinate transformation y= y( x) with a positive Jacobian, moreover, u \in C _ {2 + \alpha } ( \sigma _ {x} ) .
3) It is proved that if the coefficients of (1) belong to the space C _ \alpha ( \Omega ) and if the function u \in C _ {2+ \alpha } ( \Omega ) , then the a priori estimate
\tag{2 } \| u \| _ {C _ {2+ \alpha } ( \Omega ) } \leq C \left [ \| Lu \| _ {C _ \alpha ( \Omega ) } + \| u \| _ {C _ {2+ \alpha } ( \Omega ) } + \| u \| _ {C _ {0} ( \Omega ) } \right ]
is true up to the boundary, where the constant C depends only on \Omega , on the ellipticity constant m \leq a ^ {ij} ( x) \xi _ {i} \xi _ {j} / | \xi | ^ {2} , \xi \neq 0 , and on the norms of the coefficients of the operator L , and where
\| u \| _ {C _ {0} ( \Omega ) } = \sup _ {x \in \Omega } | u( x) | .
4) It is assumed that one knows how to prove the existence of a solution u \in C _ {2+ \alpha } to the Dirichlet problem
\left . u \right | _ \sigma = \left . \phi \right | _ \sigma ,\ \ \phi \in C _ {2+ \alpha } ( \Omega ) ,
for the Laplace operator \Delta = \sum _ {i=1} ^ {n} \partial ^ {2} / \partial x _ {i} ^ {2} .
5) Without loss of generality one may assume that \phi ( x) \equiv 0 , and then apply the continuation method, the essence of which is the following:
5 _ {1} . The operator L is imbedded in a one-parameter family of operators
L _ {t} u = tLu + ( 1- t) \Delta u ,\ \ 0 \leq t \leq 1,\ \ L _ {0} = \Delta .
5 _ {2} . Basing oneself essentially on the a priori estimate (2), it can be established that the set T of those values of t \in [ 0, 1] for which the Dirichlet problem L _ {t} u = f( x) , u \mid _ \sigma = 0 , has a solution u \in C _ {2+ \alpha } ( \Omega ) for all f \in C _ \alpha ( \Omega ) , is at the same time open and closed, and thus coincides with the unit interval [ 0, 1] .
6) It is proved that if D is a bounded domain contained in \Omega together with its closure, then for any function u \in C _ {2+ \alpha } ( D) and any compact subdomain \omega \subset D the interior a priori estimate
\tag{3 } \| u \| _ {C _ {2+ \alpha } ( \omega ) } \leq C \left [ \| Lu \| _ {C _ \alpha ( D) } + \| u \| _ {C _ {0} ( D) } \right ]
holds.
7) Approximating uniformly the functions \phi and f by functions from C _ {2+ \alpha } and applying the estimate (3), one proves the existence of a solution to the Dirichlet problem for any continuous boundary function and for a wide class of domains with non-smooth boundaries, e.g. for domains that can be represented as the union of sequences of domains \Omega _ {1} \subset \Omega _ {2} \subset \dots , with boundaries of the same smoothness as \sigma .
Estimates 2 and 3 where first obtained by J. Schauder (see [1], [2]) and go under his name. Schauder's estimates and his method have been generalized to equations and systems of higher order. The a priori estimates, both interior and up to the boundary, corresponding to it are sometimes called Schauder-type estimates. The method of a priori estimates is a further generalization of Schauder's method.
References
[1] | J. Schauder, "Ueber lineare elliptische Differentialgleichungen zweiter Ordnung" Math. Z. , 38 : 2 (1934) pp. 257–282 |
[2] | J. Schauder, "Numerische Abschätzungen in elliptischen linearen Differentialgleichungen" Studia Math. , 5 (1935) pp. 34–42 |
[3] | L. Bers, F. John, M. Schechter, "Partial differential equations" , Interscience (1964) |
[4] | R. Courant, D. Hilbert, "Methods of mathematical physics. Partial differential equations" , 2 , Interscience (1965) (Translated from German) |
[5] | A.V. Bitsadze, "Some classes of partial differential equations" , Gordon & Breach (1988) (Translated from Russian) |
[6] | Yu.M. [Yu.M. Berezanskii] Berezanskiy, "Expansion in eigenfunctions of selfadjoint operators" , Amer. Math. Soc. (1968) (Translated from Russian) |
[7] | O.A. Ladyzhenskaya, N.N. Ural'tseva, "Linear and quasilinear elliptic equations" , Acad. Press (1968) (Translated from Russian) |
Comments
Schauder-type estimates for parabolic equations were obtained for the first time in [a1] (see also [a2] for a detailed description).
References
[a1] | C. Ciliberto, "Formule di maggiorazione e teoremi di esistenza per le soluzioni delle equazioni paraboliche in due variabili" Ricerche Mat. , 3 (1954) pp. 40–75 |
[a2] | A. Friedman, "Partial differential equations of parabolic type" , Prentice-Hall (1964) |
[a3] | D. Gilbarg, N.S. Trudinger, "Elliptic partial differential equations of second order" , Springer (1977) |
Schauder method. Encyclopedia of Mathematics. URL: http://encyclopediaofmath.org/index.php?title=Schauder_method&oldid=48617