Chapter 18

Convolution

[37/30%]
Show
LaTeX
Problem 18.1

Prove that if f,g,h∈L1(R)f, g, h \in L_{1}(\mathbb {R}), then their convolution has the following properties:

?
(a)

f∗(αg+βh)=α(f∗g)+β(f∗h)f *(\alpha g+\beta h) = \alpha (f * g)+\beta (f * h) for all α,β∈C\alpha , \beta \in \mathbb {C};

(b)

(f∗g)∗h=f∗(g∗h)(f * g) * h = f *(g * h).

Problem 18.2

Prove that if f,g∈L1(R)f, g \in L_{1}(\mathbb {R}), then their convolution has the following properties:

?
(a)

f∗g∈L1(R)f * g \in L_{1}(\mathbb {R}) and ∥f∗g∥L1(R)⩽∥f∥L1(R)⋅∥g∥L1(R)\left\| f * g\right\|_{L_{1}(\mathbb {R})} \leqslant \left\| f\right\|_{L_{1}(\mathbb {R})} \cdot \left\| g\right\|_{L_{1}(\mathbb {R})};

(b)

f∗g=g∗ff * g = g * f.

Problem 18.3

Prove that supp⁡(f∗g)⊆supp⁡f+supp⁡g\operatorname {supp}(f * g) \subseteq \operatorname {supp} f+\operatorname {supp} g.

?
Problem 18.4

Compute the convolution of the characteristic functions χ[1,2](x)∗χ[3,4](x)\chi_{[1,2]}(x) * \chi_{[3,4]}(x).

?
Problem 18.5

Compute the convolution f(x)∗χ[2,4](x)f(x) * \chi_{[2,4]}(x), where

f(x)={x+1,−1⩽x⩽01−x,0<x⩽1 f(x) = \begin{cases} x+1, & -1 \leqslant x \leqslant 0 \\ 1-x, & 0 < x \leqslant 1 \end{cases}
?
Problem 18.6

Prove that the convolution ωa,b∗χ[c,d]\omega_{a, b} * \chi_{[c, d]}, where ωa,b\omega_{a, b} is the "bump" from problem 15.1, is a function of the "hat" type from problem 15.2.

?
Problem 18.7

Prove that if f,g∈L1(R),f∈C1(R)f, g \in L_{1}(\mathbb {R}), f \in C^{1}(\mathbb {R}) and f′∈L1(R)f^{\prime } \in L_{1}(\mathbb {R}), then f∗g∈C1(R)f * g \in C^{1}(\mathbb {R}) and (f∗g)′=f′∗g(f * g)^{\prime } = f^{\prime } * g.

?
Problem 18.8

Prove that if f,g∈L1(R)f, g \in L_{1}(\mathbb {R}), then f∗g^=2πf^⋅g^\widehat{f * g} = \sqrt{2 \pi } \hat{f} \cdot \hat{g}.

?
Problem 18.9

Compute the convolution 1x2+a2∗1x2+b2,a,b>0\frac{1}{x^{2}+a^{2}} * \frac{1}{x^{2}+b^{2}}, a, b > 0.

?
Problem 18.10

Prove that if f∈L1(R)f \in L_{1}(\mathbb {R}) and g∈L∞(R)g \in L_{\infty }(\mathbb {R}), then the convolution f∗gf * g is uniformly continuous on R\mathbb {R}.

?
Problem 18.11

Prove that if f∈L1(R)∩L∞(R)f \in L_{1}(\mathbb {R}) \cap L_{\infty }(\mathbb {R}) and g∈L1(R)g \in L_{1}(\mathbb {R}), then f∗g∈C0(R)∩L1(R)f * g \in C_{0}(\mathbb {R}) \cap L_{1}(\mathbb {R}) and ∥f∗g∥C0(R)⩽∥f∥L∞(R)⋅∥g∥L1(R)\left\| f * g\right\|_{C_{0}(\mathbb {R})} \leqslant \left\| f\right\|_{L_{\infty }(\mathbb {R})} \cdot \left\| g\right\|_{L_{1}(\mathbb {R})}.

?
Problem 18.12

Prove that if f∈L2(R)f \in L_{2}(\mathbb {R}) and g∈L2(R)g \in L_{2}(\mathbb {R}), then the convolution f∗gf * g is defined and belongs to the space C0(R)C_{0}(\mathbb {R}).

?
Problem 18.13

Give an example of two functions f,g∈L1(R)f, g \in L_{1}(\mathbb {R}) such that f∗gf * g is discontinuous.

?
Problem 18.14

(E. Titchmarsh, 1926). Let f,g∈L1(R),f≡g≡0f, g \in L_{1}(\mathbb {R}), f \equiv g \equiv 0 for t<0t < 0 and f∗g=0f * g = 0. Prove that then one of the functions ff or gg is zero almost everywhere.

?
Problem 18.15

Let the functions f,g∈L1(R)f, g \in L_{1}(\mathbb {R}) have support on the segment [0,1][0,1], be continuous on this segment, and f(1)≠0f(1) \neq 0. Prove that if supp⁡(f∗g)⊆[0,1]\operatorname {supp}(f * g) \subseteq [0,1], then g≡0g \equiv 0.

?
Problem 18.16

Prove that the convolution operator SfS_{f} in L2(R)L_{2}(\mathbb {R}) is unitarily equivalent to the operator of multiplication by the function 2πf^:F^SfF^−1h=2πf^h\sqrt{2 \pi } \hat{f}: \hat{F} S_{f} \hat{F}^{-1} h = \sqrt{2 \pi } \hat{f} h for every function h∈L2(R)h \in L_{2}(\mathbb {R}).

?
Problem 18.17

Find the spectrum of the convolution operator with the function e−∣x∣e^{-\left|x\right|} in L2(R)L_{2}(\mathbb {R}).

?
Problem 18.18

Find the spectrum of the operator A:L2(R)→L2(R)A: L_{2}(\mathbb {R}) \rightarrow L_{2}(\mathbb {R}) of the following form:

(Af)(x)=∫−∞+∞e−(x−t)2/2f(t)dt (A f)(x) = \int _{-\infty }^{+\infty } e^{-(x-t)^{2} / 2} f(t) d t

Is this operator compact?

?
Problem 18.19

Find the spectrum of the Hilbert operator in L2(R)L_{2}(\mathbb {R}):

Aφ(x)=1π v.p. ∫−∞+∞φ(t)x−tdt A \varphi (x) = \frac{1}{\pi } \text{ v.p. } \int _{-\infty }^{+\infty } \frac{\varphi (t)}{x-t} d t
?
Problem 18.20

(N. Wiener, 1933). Let f∈L2(R)f \in L_{2}(\mathbb {R}). Prove that the system {f(x−a):a∈R}\left\{ f(x-a): a \in \mathbb {R}\right\} of shifts of this function is complete in L2(R)L_{2}(\mathbb {R}) if and only if the measure of the set {y∈R:f^(y)=0}\left\{ y \in \mathbb {R}: \hat{f}(y) = 0\right\} equals zero.

?
Problem 18.21

Prove that for a generalized function G∈E′G \in \mathcal{E}^{\prime } the map φ→G∗φ\varphi \rightarrow G * \varphi defines a continuous operator SG:D→DS_{G}: \mathcal{D} \rightarrow \mathcal{D} and the inclusion holds:

supp⁡(G∗φ)⊂supp⁡G+supp⁡φ={x+y:x∈supp⁡G,y∈supp⁡φ}. \operatorname {supp}(G * \varphi ) \subset \operatorname {supp} G+\operatorname {supp} \varphi = \left\{ x+y: x \in \operatorname {supp} G, y \in \operatorname {supp} \varphi \right\} .
?
Problem 18.22

Prove that if a regular generalized function G=GφG = G_{\varphi } is generated by a test function φ\varphi, then for any generalized function FF the convolution F∗GF * G is a regular function, generated by the convolution F∗φF * \varphi (which is computed by Definition 18.2).

Definition 18.2: For any test function φ∈D\varphi \in \mathcal{D} (or S\mathcal{S}, or E\mathcal{E}) its convolution with a generalized function F∈D′F \in \mathcal{D}^{\prime } (respectively, S′\mathcal{S}^{\prime }, E′\mathcal{E}^{\prime }) as a function of the variable xx is defined as follows: (F∗φ)(x):=⟨F(t),φ(x−t)⟩(F * \varphi )(x) := \left\langle F(t), \varphi (x-t) \right\rangle.

?
Problem 18.23

Prove that if one of the generalized functions FF and GG is an element of E′\mathcal{E}^{\prime }, and the other an element of D′\mathcal{D}^{\prime }, then the convolution F∗GF * G is defined, is an element of the space D′\mathcal{D}^{\prime }, and the relation F∗G=G∗FF * G = G * F holds.

?
Problem 18.24

Prove that if one of the generalized functions FF and GG is an element of E′\mathcal{E}^{\prime }, and the other an element of S′\mathcal{S}^{\prime }, then the convolution F∗GF * G is defined, is an element of the space S′\mathcal{S}^{\prime }, and the relation F∗G=G∗FF * G = G * F holds.

?
Problem 18.25

Prove that if one of the generalized functions GG and HH is an element of E′\mathcal{E}^{\prime }, and the other an element of S′\mathcal{S}^{\prime }, then F^[G∗H]=2πF^[G]⋅F^[H]\hat{F}[G * H] = \sqrt{2 \pi } \hat{F}[G] \cdot \hat{F}[H].

?
Problem 18.26

Let f∈L2[−π,π]f \in L_{2}[-\pi , \pi ]. For every a∈[0,2π]a \in [0,2 \pi ] and t∈[−π,π]t \in [-\pi , \pi ] set

fa(t)={f(t−a),t−a⩾−π,f(t−a+2π),t−a<π. f_{a}(t) = \begin{cases} f(t-a), & t-a \geqslant -\pi , \\ f(t-a+2 \pi ), & t-a < \pi . \end{cases}

Prove that the system of functions {fa}a∈[0,2π]\left\{ f_{a}\right\}_{a \in [0,2 \pi ]} is complete in L2[−π,π]L_{2}[-\pi , \pi ] if and only if cn(f)=12π∫−ππf(t)e−intdt≠0c_{n}(f) = \frac{1}{2 \pi } \int_{-\pi }^{\pi } f(t) e^{-i n t} d t \neq 0 for all n∈Zn \in \mathbb {Z}.

?
Problem 18.27

Denote by D+′\mathcal{D}_{+}^{\prime } the set of generalized functions FF with the property supp⁡F⊂[0,+∞)\operatorname {supp} F \subset [0,+\infty ). Prove that if F,G∈D+′F, G \in \mathcal{D}_{+}^{\prime }, then the convolution F∗GF * G is defined and F∗G∈D+′F * G \in \mathcal{D}_{+}^{\prime }.

?
Problem 18.28

Prove that in D+′\mathcal{D}_{+}^{\prime } convolution is associative.

?
Problem 18.29

Give an example of generalized functions F,GF, G and HH for which the expressions F∗(G∗H)F *(G * H) and (F∗G)∗H(F * G) * H are defined and different.

?
Problem 18.30

Let the function g∈L1(R)g \in L_{1}(\mathbb {R}) and ∫Rgdx=1\int_{\mathbb {R}} g d x = 1. Let Fn∈D′F_{n} \in \mathcal{D}^{\prime } be the regular generalized functions generated by the functions ng(nx)n g(n x). Prove that Fn∗φ→φF_{n} * \varphi \rightarrow \varphi in D\mathcal{D} for every function φ∈D\varphi \in \mathcal{D} (cf. problems 16.20 and 18.31 а)).

?
Problem 18.31

Compute the convolutions:

?
(a)

δ0∗φ,φ∈D\delta_{0} * \varphi , \varphi \in \mathcal{D};

(b)

δ0(n)∗φ,φ∈D\delta_{0}^{(n)} * \varphi , \varphi \in \mathcal{D};

(c)

δ0(n)∗δ0(k)\delta_{0}^{(n)} * \delta_{0}^{(k)};

(d)

δ0(n)∗x+k(x)\delta_{0}^{(n)} * x_{+}^{k}(x);

(e)

(θ(x)eax)∗(θ(x)ebx)\left(\theta (x) e^{a x}\right) *\left(\theta (x) e^{b x}\right);

(f)

x+k∗x+nx_{+}^{k} * x_{+}^{n};

(g)

x+α∗x+β,Re⁡α>−1,Re⁡β>−1x_{+}^{\alpha } * x_{+}^{\beta }, \operatorname {Re} \alpha > -1, \operatorname {Re} \beta > -1.

Problem 18.32

Show that for a linear continuous operator A:D→EA: \mathcal{D} \rightarrow \mathcal{E} the following properties are equivalent:

  1. AA commutes with shifts: (Aφ)(x−a)=A(φ(x−a))(A \varphi )(x-a) = A(\varphi (x-a)) for all a∈Ra \in \mathbb {R};

  2. А commutes with differentiation: (Aφ)′=A(φ′)(A \varphi )^{\prime } = A\left(\varphi^{\prime }\right);

  3. Aφ=F∗φA \varphi = F * \varphi for all φ\varphi, where F∈D′F \in \mathcal{D}^{\prime } is some generalized function.

?
Problem 18.33

Compute the antiderivative of order 1/21 / 2 of

?
(a)

θ(x)\theta (x);

(b)

χ[0,1](x)\chi_{[0,1]}(x).

Problem 18.34

Verify that I(12)I(12)θ(x)=x+I\left(\frac{1}{2}\right) I\left(\frac{1}{2}\right) \theta (x) = x_{+}.

?
Problem 18.35

Prove that I(α)I(β)f=I(α+β)fI(\alpha ) I(\beta ) f = I(\alpha +\beta ) f for f∈D+′,α,β∈(−12,12)f \in \mathcal{D}_{+}^{\prime }, \alpha , \beta \in \left(-\frac{1}{2}, \frac{1}{2}\right).

?
Problem 18.36

Compute the derivative of order 1/21 / 2 of θ(x)\theta (x).

?
Problem 18.37

Prove that the Abel integral equation

g(x)=1Γ(1−α)∫0xf(t)dt(x−t)α g(x) = \frac{1}{\Gamma (1-\alpha )} \int _{0}^{x} \frac{f(t) d t}{(x-t)^{\alpha }}

with known differentiable function gg has the solution f(x)=1Γ(α)∫0xg′(t)dt(x−t)1−αf(x) = \frac{1}{\Gamma (\alpha )} \int_{0}^{x} \frac{g^{\prime }(t) d t}{(x-t)^{1-\alpha }} for every α∈(0,1)\alpha \in (0,1).

?