Научная статья на тему 'About the equality of the transform of Laplace to the transform of Fourier'

About the equality of the transform of Laplace to the transform of Fourier Текст научной статьи по специальности «Математика»

CC BY
112
31
i Надоели баннеры? Вы всегда можете отключить рекламу.
Журнал
Проблемы анализа
WOS
Scopus
ВАК
MathSciNet
Область наук
Ключевые слова
LAPLACE TRANSFORM / FOURIER TRANSFORMS / DIRICHLET PROBLEM / NEW INVERSE OF LAPLACE TRANSFORM

Аннотация научной статьи по математике, автор научной работы — Pavlov A.V.

We proved that the transform of Laplace does not have complex part on the complex axis for the wide class of functions in different situations. The main theorem is proved presenting a function as sum of two Laplace transforms. The transforms are defined in the left and right parts of the plain accordingly. Such presentation is proved to be unique. With help of the results we obtain equality of the transforms of Laplace and Fourier for some class of functions.

i Надоели баннеры? Вы всегда можете отключить рекламу.
iНе можете найти то, что вам нужно? Попробуйте сервис подбора литературы.
i Надоели баннеры? Вы всегда можете отключить рекламу.

Текст научной работы на тему «About the equality of the transform of Laplace to the transform of Fourier»

DOI: 10.15393/j3.art.2016.3211

UDC 517.442, 517.443, 517.547.3

A. V. Pavlov

ABOUT THE EQUALITY OF THE TRANSFORM OF LAPLACE TO THE TRANSFORM OF FOURIER

Abstract. We proved that the transform of Laplace does not have complex part on the complex axis for the wide class of functions in different situations. The main theorem is proved presenting a function as sum of two Laplace transforms. The transforms are defined in the left and right parts of the plain accordingly. Such presentation is proved to be unique. With help of the results we obtain equality of the transforms of Laplace and Fourier for some class of functions.

Key words: Laplace transform, Fourier transforms, Dirichlet problem, new inverse of Laplace transform

2010 Mathematical Subject Classification: 44A10, 44A38 1. Introduction. We consider the transform of Laplace in the form

LF_u(x)(-)(v), v E [0, +to),

where, by definition,

to

e

vx

0

to

J e±ixy u(x)dx = F±u(x)()(y), y E (-to, to).

—to

The basic result of the article is formulated in Theorem 2:

(n/2)u(t) = LC0u(x)O(t), t E (0, to),

©Petrozavodsk State University, 2016

if u(-x) = u(x), x E (—ro, ro) (in the condition of Theorem 2), where

CO

/cosvxS(x)dx = C»S(*)(•)(„), v E (—ro,ro).

0

Similar results are considered in works [1] - [6] in connection with the problem of the inverse of transform of Laplace ([2], [3], [4]). Traditional methods, related to regular functions ([1], [3], [6]-[9]), are not applicable to functions u(t) = u2 (x) without properties u2 (—x) = u2 (x) or u2 (—x) = = —u2(x). The direction is considered in the second part of proof of Theorem 1.

The main result (Theorem 1) follows from presenting function u(ip) as a sum of Laplace transforms

L(P) + (—L1(P)) = u(iP),

(the exact formulation is in Lemma 1), if the transforms are defined in the left and right parts of the complex plain accordingly, where, by definition,

<X>

L1 (p) = (1/2n) / u(x)/(p + ix)dx, Rep < 0,

L(p) = (1/2n) J u(x)/(p + ix)dx, Rep > 0.

With help of this presentation we obtain the equality of the transform of Laplace to the transform of Fourier (Theorem 2).

In general all results follow from the Re L(is) = L(is) equality of Theorem 1, if s E (—ro, ro).

The results of Theorem 1 are related to the problem of Dirichlet and the integrals of Poisson and Schwartz [7, p. 209]. This subject is studied in [1], [4]; we do not touch it here.

The proof follows from some main facts: in rather wide conditions we can use (for s E (—ro, ro))

2nL(—is) = y eisxdx j e—txu(t)dt = 0

CO CO CO CO

= J cos sxdx J 2 cos txu(t)dt + ij sin sxdx J 2 cos txu(t)dt =

0 0 0 0 C CO

= nu(s) + 2i J sin sxdx J cos txu(t)dt, u(-1) = u(t); 00

CO C

2nL(—is) = nu(s) — 2i j cos sxdx J sin txu(t)dt, u(-t) = -u(t),

00

and we obtain the sum (Lemma 1)

L(is) + (—Li(is)) = u(—s), s E (—to, to), L(p) + (—Li(p)) = u(ip),

ip E J, where function L(p) is defined for Re p > 0, and function L1(p) with the same analytical expression is defined for Rep < 0, J is a set of the special points of function u(p); we use equalities

L(—is) = L(is), L(is) = —L1(is)), u(—s) = u(s), s E (—to, to).

In the main second part of proof of Theorem 1 we consider function u2(p) = L(p) in the form

l(is) + l(—is) = u2(—s) = L*(is) + L^(is), s E (—to, 0),

where

C

L* (p) = (1/2n) J u2(x)/(p + ix)dx, Rep > 0,

— C

L(x) = u2(x) = Reu2(x), x E (—to, +to), and function L*(p) with the same analytical expression is defined for Re p < 0, where, by definition,

l(p) = (1/2n)LC02u2(x)(-)(p), Rep > 0,

but l(is) + l(—is) = l(—is) + l(—is), and we get

L(is) = u2(is) = u2(—is) = L(—is), s E (—to, +to),

or (Theorem 1) u(s)/2 = Reu2(is) = u2(is) = L(is) , s E (—to, +to).

The main part of the article is the second part of proof of Theorem 1.

2. The transform of Laplace on the complex axis. By definition, the area of regularity of function u(p) is Gu.

Theorem 1. If u(-p) = u(p), p G Gu, u(p) is regular in the C plane, except a finite number of points J = (zi,... , zk, Re zk = 0, Im zk = = 0, k > 2}, and

u(0) = 0, |u(s)| < c/|s|2, |s| —> to, s G (-to, to), c < to,

where c is the constant, the equality

CC

u(-y)/2 = lim (1/2n) / u(t)/((x + iy) + it)dt, y G (-to, to),

J

takes place, and the equality

nu(it) = LF— u(x)()(t), t G (-to, to),

takes place if

lim (1/2n)LF— u(t)(-)(x + iy) = (1/2n)LF— u(t)()(iy), y G (-to, to).

Proof. We use Lemma 1 (a part of Lemma 1 was proved in [2]-[5]. In the Appendix we give a new full proof).

Lemma 1. 1. Equality L(is) - L1 (is) = u(-s), s G (-to, to), takes place if the function u(s) is continuous for all s G (-to, to); equality

L(p) - Li(p) = u(ip), ip G Gu,

takes place if u(p) is regular in (p : |Imp| < a} G Gu, for a constant a G (0, +to). c

2. lim 2nL(p) = 2nL(iy) = nu(-y) - / (u(x - y)/ix)dx,

p^iy, Re p>0 J

— oo

lim = ^Lii-iy)' Li (zy) = -L(iУ),

y G (-to, to), in all points y : du(s)/ds\s=y < to, if the function u(s) is continuous for all s G (-to, to), Re u(s) = u(s), s G (-to, to), and

\u(s)\ < c/\s\5, s ^ ±to, s G (-to, to), 5 > 0, 5 = const.

3. If u(0) = 0, and du(s)/ds|s=0+ < to, du(s)/ds|s=0— < to values are defined, but du(s)/ds|s=0+ = du(s)/ds|s=0—, the second part of Lemma 1 takes place for y = (—to, to) (and for y = 0).

From Lemma 1 in the conditions of Theorem 1 we obtain u(ip) = L(p) — L1 (p), ip E J.

We will assume that

u(ip) = ui(p) + u2 (p) = L(p) — Li(p), ip E J,

where u1 (p), u2(p) are new functions: u1(p) is regular in the left part of the complex plane (without points Rep = 0 ), u1(p) is bounded in the left part of the complex plane (with all points Rep = 0), and u1(p) is continuous in all points on the complex axis (from the left part of the plain); u2(p) is regular in the right part of the complex plane (without points Re p = 0), u2(p) is bounded in the right part of the complex plane (with all points Rep = 0), and u2(p) is continuous in all points on the complex axis (from the right part of the plain).

We get

u2(p) = L(p^ u1(p) = — UCp^ — l1(p) = L(—p^ p E J;

the fact we obtain from

u2 (p) — L(p) = —u1 (p) + (—L1 (p)) = 0,

where the functionin the left-hand side has special points in the left part of the complex plane (see Lemma 1), and the function in the right-hand side has special points in the right part of the complex plane (see Lemma 1), and both functions are equal to 0 in to (it is obvious from Lemma 1). By definition u2(p) — L(p) is regular in the right part of the plain (with Rep = 0), and —u1 (p) + (—L1(p)) is regular in the left part of the plain (with Rep = 0). We obtain that both functions are regular in all points of the complex plain, and both functions are bounded by a constant in their domains (this is obvious from the definition), or both functions are zero ([7]).

As a result we have Re u2(s) = u2(s), s E (—to, to), where u2(—s) = L1(—s) + u(—is) = —L(s) + u(—is), u(is) = u(—is), s E [0, to).

The second part of the proof of Theorem 1. We will prove that Reu2(is) = u2(is),s E (—to, to), where u2(p) = L(p).

With help of Lemma 1 we will repeat the previous reasoning for new pairs of functions (function u2 (x) can be neither odd nor even)

2nL*(p) = J u2(x)/(p + ix)dx, Rep > 0,

—c CC

2nL*(p) = J u2(x)/(p + ix)dx, Rep < 0.

—C

We get

L*(is) + (-L*(is)) = U2(-s), s G (-to, to), L*(p) + (-Ll(p)) = U2(ip),

ip = J, (in the equality we use Lemma 1 with u2 (x) instead of function u(x), Reu2(x) = u2(x), x G (-to, to) ). We use the new function

u0(x) = u2(|x|), x G (-to, to).

With help of Lemma 1 we get

iНе можете найти то, что вам нужно? Попробуйте сервис подбора литературы.

l(is) + li (is) = U2 (|s|) = U2 ( s), s G (-to, 0),

(not for s G (-to, to)), where

2n1(p) = J u2(|x|)/(p +ix)dx Rep > 0,

—C

2n1i(p) = -/u2(lxl)/(p +ix)dx Rep <0.

For all positive -s we obtain : U2 (-s) = L* (is) + (-L*(is)) = l(is) + li (is) = U2(|s|), -s G (0, +to),

and (as in Lemma 1, but only for positive -s G (0, +to)) we get ([7]) U2 (ip) = L*(p) + (-Ll(p)) = 1(p) + h (p),

in the area of joint regularity of the sum u2(ip), p G J• (We use the theorem about analytical continuation across a line ([7]), where functions L* (p), L*(p), l(p), li(p) are continuous on the complex axis in their domains).

But 11 (p) = 1(-p), Rep < 0 (it is obvious from the values on (-to, 0)), and we get

l(p) + li(p) = l(p) + l(-p) = u2 (ip) = L* (p) + (-L*(p)), ip = J.

From the equality

l(p) - L*(p) = -l(-p) + (-Ll(p)), ip = J,

with help of continuity of functions l(p), 11(p) = 1(-p) on the complex axis and regularity of function -L*(p), L*(p) for all ip = J, p G (—iTO, iTO) we obtain that

l(p) = L*(p) - l(-p) + (-L1 (p)), Rep < 0

is the analytical continuation of function l(p) across all points of the complex axis from the right part of the plain to the left part of the plain (theorem about analytical continuation across the line [7]).

We have proved u2 (-pi) = u2(pi) = l(p) + 1(-p), ip G J, and

Reu2(is) = u2(is) = L(is) = Re L(is), s G (-to, to)

(for the function u(-p) = u(p) ([7])). We can see

u(-y)/2 = Re L(iy) = L(iy) = lim L(x + iy), y G то),

(from Lemma 1), and (changing limits of integration ([10]))

u(-y)/2 = (1/2n)LF_u(t)(-)(iy) = L(iy), u(-y) = u(y), y G (-to, to),

if lim (1/2n)LF_u(t)(0(x + iy) = (1/2n)LF_u(t)(^)(iy), y G (-to, to). Theorem 1 is proved. □

From Theorem 1 and formula

LF-u(x)(0(it) = 2LC0u(x)(-)(t), t E (-to, to), we obtain Theorem 2.

Theorem 2. In the conditions of Theorem 1

nu(t) = 2LC0u(x)(0(t), t E (0, to), if u(p) = f (p4), p E J, where function f (z) is regular for all z = p4, p E J-

3. Appendix. Proof of Lemma 1. From [10]

2n(L(x + iy) — Li(-x + iy)) =

CC CC

= ^ [u(t)/(x + iy + it)]dt — J [u(t)/(—x + iy + it)]dt =

-C -C

= ^ [u(t1 — y)2x/(x2 + ti)]dt1 = 2 J [u(t2x — y)/(1+ t|)]dt2 ^ 2nu(—y),

x ^ 0, y E (—to, to).

The first part of Lemma 1 is proved. From [10]

2nL(x + iy) = J [u(t)/(x + iy + it)]dt = J [u(t1 — y)/(x + it1 )]dt1 =

—C —C

CC CC

= [u(t1 — y)x/(x2 + t1 )]dt1 — i [u(t1 — y)t1 /(x2 + t2 )]dt1 =

= J [u(xt2 — y)/(1+ t2)]dt2 —

—C

CC CC

—i (u(t1 — y)/t1)dt1 + i [(u(t1 — y)/t1 )x2/(x2 + t1 )]dt1 ^

^ nu(-y) - i (u(ti - y)/ti)dti,y E (-TO, to);

we use

r (u(ti - y)/ti)xd _ I ((u(ti - y) - u(-tl - y))/tl)xdt _

i (x2 + tl) dtl i (x2 +12) dtl

—C —C

CO

_ x J [2U(tl - y - 20tl)x/(x2 + tl)]dtl _

—C

_ x J [2U(xt2 -y-20lxt2)/(1+12)]dt2 _ xn(U(-y) + o(1)) ^ 0, x ^ 0,

—C

2U(-y) < to, y E (-to, to), 0 < 0 < 1, 0 < 0l < 1,

where we use notation du(t)/dt _ U(t), (if du(t)/dt _ U(t) is defined for all t E (-to, to)), and the second part of Lemma 1 is proved.

If u(0) _ 0, and du(s)/ds|s=0+, du(s)/ds|s=0—, s E (-to, to), values are defined, we use for y = 0

c 0

J [(u(tl)/tl)x2/(x2 + tl)]dtl _ x J [([u(tl) - u(0)]/tl)x/(x2 + tl)]dtl +

—C —C

+ ^ l(Ktl) - u(0)]/tl)x/(x2 + il)ldil _A(x) ^ «,* ^ 0;

o

Im2nL(x + i0) _ - J (u(tl - 0)/tl)dtl + A(x) ^

—C

^ - J ([u(tl) - u(0)]/tl)dtl, x ^ 0.

—C

For Ll(-x + iy) ^ Ll(iy), -x _ Rep < 0, the equalities are proved in the similar way. □

References

[1] Pavlov A. V. The integrals of Poisson and Schwartz and the transformation of Laplace. Intern. Jour. of Information Research and Review, 2015, vol. 2, no. 2, pp. 394-400.

[2] Pavlov A. V. A new inversion formula for Laplace transforms and the notion of evenness. Univer. Jour. of Appl. Math., Hor. Resear. Publ., 2014, vol. 2, no. 3, pp. 148-152. DOI: 10.13189/ujam.2014.020305.

[3] Pavlov A. V. The new inversion of Laplace transform. Jour. of Math. and Syst. Scien., David Pub. Sciences, 2014, vol. 4, no. 3, pp. 197-201.

[4] Pavlov A. V. The Fourier, Laplace transformations and the Newton potential. Amer. Jour. of Appl. Mathem. and Stat., Scien. and Educ. Publ., 2015, vol. 2, no. 6, pp. 398-401. D0I:10.12691/ajams-2-6-7.

[5] Pavlov A. V. Reliable prognosis of the functions in the form of transformations of Fourier or Laplace. Herald of MIREA, Mosc. Inst. of Rad., Electr. and Autom. (Mosc. Tech. Univ.), 2014, vol. 3, no. 2, pp. 78-85.

[6] Pavlov A. V. The Fourier transform and the formula of Laplace transforms. Mathem. Notes, Springer, 2011, vol. 90, no. 6, pp. 792-796.

[7] Lavrentiev M. A., Shabat B. V. The methods of theory of functions of complex variable. Science, Moscow, 1987, 688 p.

[8] Davis B. J. Integral transforms and their applications (3 eddition). Springer Verlag, New York, 2002.

[9] Kolmogorov A. N., Fomin S. V. Elements of the theory of functions and functional analysis. Moscow: Science, Moscow, 1976, 544 p.

[10] Fichtenholz G. M. Course of differential and integral calculus. Moscow.: Science, Moscow, vol. II, 1969, 800 p.

Received February 06, 2016.

In revised form, July 05, 2016.

Accepted June 30, 2016.

Moscow Institute of Radio-technics, Electronics and Automatics (MTU) 78, Vernadsky Avenue, 119454 Moscow, Russia E-mail: a_pavlov@mirea.ru

i Надоели баннеры? Вы всегда можете отключить рекламу.