Problem set 6

Problem 1

(1) Prove the last lemma in lecture 12.

(2) Conclude that the Fourier transform F:SS\mathcal F: \mathcal S\to \mathcal S, fFff\mapsto \mathcal Ff is continuous with respect to the topology on S\mathcal S.

Problem 2

Let K(x)=eπx2K(x) = e^{-\pi x^2} be the Gaussian function. Similar to what we did with Λ and Λϵ\Lambda_\epsilon, we define Kδ(x)=1δeπx2δK_\delta(x) = \frac{1}{\sqrt \delta}e^{-\frac{\pi x^2}{\delta}}.

(1) Show that KδK_\delta is a good kernel on R\mathbb{R} when δ0\delta \to 0, in the sense that +Kδ(x)dx=1;+Kδ(x)dxM\int_{-\infty}^{+\infty}K_\delta(x)dx = 1; \int_{-\infty}^{+\infty}|K_\delta(x)|dx \leq M; and for η>0\eta> 0, x>ηKδ(x)dx0\int_{|x|>\eta}|K_\delta(x)|dx \to 0 when δ0\delta \to 0.

(2) If ff is continuous and bounded on R\mathbb{R}, then fKδff*K_\delta \to f uniformly when δ0\delta \to 0.

(3) Calculate the Fourier transform of KδK_\delta.

(4) Let TT be an operator on SS given by (T,f):=limδ0(TKδ,f)(T,f):=\lim_{\delta\to 0} (T_{K_\delta},f), calculate TT.

(5) Show that f(0)=+Ff(s)dsf(0) = \int_{-\infty}^{+\infty}\mathcal Ff(s)ds for fSf\in \mathcal S.

(6) Give an alternative proof of inversion theorem by looking at F(y)=f(y+x)F(y) = f(y+x).

Problem 3

(1) For k0k\geq 0, show that for every fCck(R)f\in \mathscr{C}_c^k(\mathbb{R}), there exists a sequence fnf_n in Cc(R)\mathscr{C}_c^{\infty}(\mathbb{R}) such that fnff_n\to f in Cck(R)\mathscr{C}_c^k(\mathbb{R}).

(2) Let TT be a distribution of order kk. Show that uu extends uniquely to a continuous linear functional on Cck(R)\mathscr{C}_c^k(\mathbb{R}). In other words, define T(φ)T(\varphi) for every φCKk(R)\varphi\in \mathscr{C}_K^k(\mathbb{R}) and show that (T,φ)C0lkφl,K|(T,\varphi)|\leq C\sum_{0\leq l\leq k}\|\varphi\|_{l,K}, where CC is a constant depending on KK.

Problem 4 (optional)

(1) Let T:SST:\mathcal S\to \mathcal S be a linear operator. Show that if TT commutes with both ddx\frac{d}{dx} and 2πix2\pi i x, i.e. Tddxf=ddxTfT\frac{d}{dx}f = \frac{d}{dx}Tf, T(xf)=x(Tf)T(xf) = x(Tf), then T=cIT = cI for some constant cc, i.e. Tf=cfTf = cf for every fSf\in S.

(2) Let T:f(F2f)T:f\mapsto (\mathcal F^2f)^-. Show that TT satisfies (1) and determine cc for TT.

(3) Conclude that the Fourier transform F:SS\mathcal F: \mathcal S\to S is an isomorphism without directly using the Fourier inversion theorem. (This is an alternative proof of the inversion theorem.)

Problem 5

Prove that fSf\in S and TST\in S' then fTf*T is a tempered distribution.