292 CHAPTER 11. FUNDAMENTAL TRANSFORMS

subspace even though this n− 1 dimensional subspace has measure zero, wheneverof u ∈ H1 (Rn). Hint: u(x′,0) =

∫∞

0 Dn

(−e−s2

u(x′,s))

ds =∫

0 −2se−s2u(x′,s)−

e−s2Dnu(x′,s)ds.

Thus ∣∣γu(x′)∣∣2 ≤C

(∫∞

02se−s2 ∣∣u(x′,s)∣∣2 ds+

∫∞

0e−s2 ∣∣Dnu

(x′,s)∣∣2 ds

).

Now do∫Rn−1 to both sides.

15. For u ∈ G , let γu(x′) ≡ u(x′,0) . Justify the following arguments. F ′ refers to theFourier transform with respect to x′, (x1, ...,xn−1). This and the next problem are ona more refined version of Problem 14.∫

RFu(x′,xn

)dxn = lim

ε→0

∫R

e−(εxn)2Fu(x′,xn

)dxn

= limε→0

(1

)n/2 ∫Rn

u(y′,yn

)e−ix′·y′

∫R

e−(εxn)2e−ixnyndxndy′dyn

= limε→0

Kn

∫Rn

u(y′,yn

)e−ix′·y′e−ε2 y2

n4

∫R

e−ε2(

xn+iyn2

)2

dxndy′dyn

= limε→0

Kn

∫Rn

u(y′,yn

)e−ix′·y′e−ε2 y2

n4

dy′dyn

= K̂n

∫Rn

u(y′,0

)e−ix′·y′dy′

F ′ (γu)(x′)=Cn

∫R

Fu(x′,xn

)dxn.

16. ↑First show that if a > 0 and t > 1/2, then∫R(a2 + x2

)−t dx <Cta1−2t . Next explainthe following steps where Kn is a constant depending on n∫

Rn−1

(1+∣∣y′∣∣2)s ∣∣Fγu

(y′)∣∣2 dy′

= Cn

∫Rn−1

(1+∣∣y′∣∣2)s

∣∣∣∣∫RFu(y′,yn

)dyn

∣∣∣∣2 dy′

=Cn

∫Rn−1

(1+∣∣y′∣∣2)s

∣∣∣∣∫RFu(y′,yn

)(1+ |y|2

)t/2(1+ |y|2

)−t/2dyn

∣∣∣∣2 dy′

Now apply the Cauchy Schwarz inequality to get:

≤ Cn

∫Rn−1

(1+∣∣y′∣∣2)s ∫

R

∣∣Fu(y′,yn

)∣∣2(1+ |y|2)t

dyn

·∫R

(1+ |y|2

)−tdyndy′.

Now use the first part with a2 = 1+ |y′|2. Obtain

≤ Cn

∫Rn−1

(1+∣∣y′∣∣2)s(

1+∣∣y′∣∣2)(1−2t)/2

·∫R

∣∣Fu(y′,yn

)∣∣2(1+ |y|2)t

dyndy′.

29215.16.CHAPTER 11. FUNDAMENTAL TRANSFORMSsubspace even though this n — | dimensional subspace has measure zero, wheneverof wu € H'(R"). Hint: u(x’,0) = fy Dn (-e"u(x',s)) ds= fy —2se~"u(x',s) -e-* Dau (x’,s) ds.Thus| yu (x! Pse( 2se7 | lu(x'.s) Pas [oe * Duu(x's) Pas).Now do fgn-1 to both sides.For u € Y, let yu(x’) = u(x’,0). Justify the following arguments. F’ refers to theFourier transform with respect to x’, (x1,...,%,—1). This and the next problem are ona more refined version of Problem 14.2[Fa x’ Xn) ) dx, = =lim | e~@*)" Fu (x’, Xn) dXp€-0 Re1 n/2 poy >,_ | I) poixly | —(€%n)° g®nYn dy, dy'des) [ue nde _ ° een2= lim Ky fu u(y’. yn) ey OE fee °C ) dxady'dy,R2A= lim Ky uly Yn) ey eet i_ R, [us eo ¥ dy!F' (yu) (x') = Cr ff Fu(x',xn) dXp.R+ First show that if a > 0 and > 1/2, then fg (a? +x’) dx <C,a!~*. Next explainthe following steps where K,, is a constant depending on nto (1+ |y'?) [Fre(y’) Pay’- Gf (1+?) | [ru (y'.yn) dyn=G.[ (1 + ly?) [ee (y’.yn) (1 +lyP)"" (1 +lyP) ay,Now apply the Cauchy Schwarz inequality to get:t< Gf (1+ly/) [|Fu(on)? (+l?) don| (14191?) “aynay’.RNow use the first part with a2 = 1 + |y’|?. Obtain2dy’2dy’< G on (1+|y'P) (1+ iP"t[Fu (y'.yn) |? (1 +19!) dyndy’.