Summary
In Section 1.3 we defined the convolution u1*u2 of two continuous functions u1 and u2, one of which has compact support. The definition can be applied without change if u1∈D’, (resp. D’,) and u2∈C0∞ (resp. C∞) we have u1*u2òC∞ then. Section 4.1 is devoted to such convolutions. As in the case of functions this is an efficient method to approximate distributions by C∞ functions. It can often be used to extend statements concerning smooth functions to distributions, particularly when translation invariant questions are concerned. As examples of this we give in Section 4.1 a discussion of convex, sub-harmonic and plurisubharmonic functions.
Convolution of two distributions u1 u2 one of which has compact support is defined in Section 4.2 so that the associativity (u1 * u2)* ϕ = u1 *(u2 * ϕ), ϕ∈C0∞, is preserved. It is then elementary to see that supp(u1 *u2)⊂:suppu1+suppu2.
Section 4.3 is devoted to the proof of the theorem of supports which states that when u1 and u2 both have compact supports, then there is equality if one takes convex hulls of the supports. The standard proofs of this depend on analytic function theory, and we shall return to them later on. The reader might therefore prefer to wait for Section 7.3 rather than studying the end of the proof of Theorem 4.3.3.
Section 4.4 is intended to present the basic methods used to derive results on existence and smoothness of solutions of constant coefficient partial differential equations from the properties of a fundamental solution. This is an important application of the convolution. The final Section 4.5 is then devoted to Lp estimates for convolutions. In addition to estimates related to Hölder’s inequality we prove potential estimates of the Hardy-Littlewood-Sobolev type and derive from them relations between the Lp (or Hölder) classes of a function and its derivatives. These are basic particularly in the study of elliptic differential equations. They will be supplemented in Section 7.9 when we have Fourier analysis at our disposal.
Preview
Unable to display preview. Download preview PDF.
Rights and permissions
Copyright information
© 1998 Springer-Verlag New York, Inc.
About this chapter
Cite this chapter
Hörmander, L. (1998). Convolution. In: The Analysis of Linear Partial Differential Operators I. Grundlehren der mathematischen Wissenschaften, vol 256. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-96750-4_5
Download citation
DOI: https://doi.org/10.1007/978-3-642-96750-4_5
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-96752-8
Online ISBN: 978-3-642-96750-4
eBook Packages: Springer Book Archive