Integral Transforms with Non-separated Variables and Discontinuous Coefficients

Multidimensional integral transforms with non-separated variables for discontinuous coefficients problems are constructed in the case where the coefficient discontinuities are on the parallel hyperplanes. Explicit kernel formulas for ideal coupling conditions are obtained. The basic integral transfo...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Журнал математической физики, анализа, геометрии
Datum:2013
1. Verfasser: Yaremko, O.E.
Format: Artikel
Sprache:English
Veröffentlicht: Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України 2013
Online Zugang:https://nasplib.isofts.kiev.ua/handle/123456789/106772
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Назва журналу:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Zitieren:Integral Transforms with Non-separated Variables and Discontinuous Coefficients / O.E. Yaremko // Журнал математической физики, анализа, геометрии. — 2013. — Т. 9, № 4. — С. 594-603. — Бібліогр.: 10 назв. — англ.

Institution

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id nasplib_isofts_kiev_ua-123456789-106772
record_format dspace
spelling Yaremko, O.E.
2016-10-04T20:06:59Z
2016-10-04T20:06:59Z
2013
Integral Transforms with Non-separated Variables and Discontinuous Coefficients / O.E. Yaremko // Журнал математической физики, анализа, геометрии. — 2013. — Т. 9, № 4. — С. 594-603. — Бібліогр.: 10 назв. — англ.
1812-9471
https://nasplib.isofts.kiev.ua/handle/123456789/106772
Multidimensional integral transforms with non-separated variables for discontinuous coefficients problems are constructed in the case where the coefficient discontinuities are on the parallel hyperplanes. Explicit kernel formulas for ideal coupling conditions are obtained. The basic integral transform identity is proved.
Сконструированы многомерные интегральные преобразования с неразделенными переменными для задач с разрывными коэффициентами в случае, когда разрывы коэффициентов сосредоточены на параллельных гиперплоскостях. Найдены явные формулы для ядер для случая условий идеального сопряжения. Доказано основное тождество интегрального преобразования.
en
Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України
Журнал математической физики, анализа, геометрии
Integral Transforms with Non-separated Variables and Discontinuous Coefficients
Article
published earlier
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
title Integral Transforms with Non-separated Variables and Discontinuous Coefficients
spellingShingle Integral Transforms with Non-separated Variables and Discontinuous Coefficients
Yaremko, O.E.
title_short Integral Transforms with Non-separated Variables and Discontinuous Coefficients
title_full Integral Transforms with Non-separated Variables and Discontinuous Coefficients
title_fullStr Integral Transforms with Non-separated Variables and Discontinuous Coefficients
title_full_unstemmed Integral Transforms with Non-separated Variables and Discontinuous Coefficients
title_sort integral transforms with non-separated variables and discontinuous coefficients
author Yaremko, O.E.
author_facet Yaremko, O.E.
publishDate 2013
language English
container_title Журнал математической физики, анализа, геометрии
publisher Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України
format Article
description Multidimensional integral transforms with non-separated variables for discontinuous coefficients problems are constructed in the case where the coefficient discontinuities are on the parallel hyperplanes. Explicit kernel formulas for ideal coupling conditions are obtained. The basic integral transform identity is proved. Сконструированы многомерные интегральные преобразования с неразделенными переменными для задач с разрывными коэффициентами в случае, когда разрывы коэффициентов сосредоточены на параллельных гиперплоскостях. Найдены явные формулы для ядер для случая условий идеального сопряжения. Доказано основное тождество интегрального преобразования.
issn 1812-9471
url https://nasplib.isofts.kiev.ua/handle/123456789/106772
citation_txt Integral Transforms with Non-separated Variables and Discontinuous Coefficients / O.E. Yaremko // Журнал математической физики, анализа, геометрии. — 2013. — Т. 9, № 4. — С. 594-603. — Бібліогр.: 10 назв. — англ.
work_keys_str_mv AT yaremkooe integraltransformswithnonseparatedvariablesanddiscontinuouscoefficients
first_indexed 2025-11-25T20:40:19Z
last_indexed 2025-11-25T20:40:19Z
_version_ 1850530674699665408
fulltext Journal of Mathematical Physics, Analysis, Geometry 2013, vol. 9, No. 4, pp. 594–603 Integral Transforms with Non-separated Variables and Discontinuous Coefficients O.E. Yaremko Penza State University 37, Lermontov Str., Penza 440038, Russia E-mail: yaremki@mail.ru Received March 21, 2012, revised January 21, 2013 Multidimensional integral transforms with non-separated variables for discontinuous coefficients problems are constructed in the case where the coefficient discontinuities are on the parallel hyperplanes. Explicit kernel formulas for ideal coupling conditions are obtained. The basic integral trans- form identity is proved. Key words: integral transforms, non-separated variables, coupling con- ditions. Mathematics Subject Classification 2010: 65R10. 1. Introduction It is known that the structure of integral transforms, which can be used to solve the boundary value problems, is determined by the type of differential equation. A number of transforms have appeared in mathematical literature since the 70th of the last century in the works by Y.S. Uflyand [1, 2], M.P. Lenuk [3, 4], L.S. Nayda [4], V.S. Protsenko [5], etc. In particular, the author and I.I. Bavrin [7] have proposed integral transforms with non-separated variables for solving multidimensional problems. Let V ⊂ Rn+1 be the half-space V = { (y1, . . . , yn, x) ∈ Rn+1 : x > 0 } . Then the solution of the Dirichlet problem is expressed via the Poisson formula [8], u(x, y) = Γ ( n + 1 2 ) π− n+1 2 ∫ y=0 x [ (y − η)2 + x2 ]n+1 2 f(η)dη. c© O.E. Yaremko, 2013 Integral Transforms with Non-separated Variables Obviously the Poisson kernel has the form of the Laplace transform Γ ( n + 1 2 ) π− n+1 2 x [ x2 + (y − η)2 ]n+1 2 = ( 1 2π )n 2 ∫ ∞ 0 λ n 2 e−λx Jn−1 2 (λ |y − η|) |y − η|n−2 2 dλ, where Jν is the Bessel function of order ν [8]. Reproducing property of the Poisson kernel is obtained from the expansion of the function f(y) with respect to the Laplace operator ∆ eigenfunctions: f(y) = lim τ→0 ∞∫ 0 λ n 2 e−λτ   1(√ 2π )n ∫ Rn Jn−2 2 (λ |y − η|) |y − η| n−2 2 f (η) dη   dλ. On the basis of this expansion we can conclude that the integral transforms with non-separated variables are defined as follows [7]: direct integral Fourier transform F [f ] (y, λ) = 1(√ 2π )n ∫ Rn Jn−2 2 (λ |y − η|) |y − η| n−2 2 f (η) dη ≡ f̂ (y, λ) , (1) inverse Fourier integral transform F−1[f̂ ](y) = lim τ→0 ∞∫ 0 λ n 2 e−λτ f̂(y;λ)dλ ≡ f(y). (2) The goal of the paper is to construct multi-dimensional analogues of integral transforms (1), (2) appropriate for the differential equations with discontinuous coefficients. 2. One-dimensional Integral Transforms with Discontinuous Coefficients In the paper, the integral transforms with discontinuous coefficients are con- structed in accordance with author’s work [10]. Let ϕ (x, λ) and ϕ∗ (x, λ) be eigenfunctions of the direct and the dual Sturm–Liouville problems for the Fourier operator on piecewise-homogeneous axis In, In = { x : x ∈ n+1 U j=1 (lj−1, lj) , l0 = −∞, ln+1 = ∞, lj < lj+1, j = 1, n } . Journal of Mathematical Physics, Analysis, Geometry, 2013, vol. 9, No. 4 595 O.E. Yaremko The eigenfunction ϕ (x, λ), ϕ (x, λ) = n∑ k=2 θ (x− lk−1) θ (lk − x) ϕk (x, λ) + θ (l1 − x) ϕ1 (x, λ) + θ (x− ln) ϕn+1 (x, λ) , is the solution of the system of separated differential equations ( a2 m d2 dx2 + λ2 ) ϕm (x, λ) = 0, x ∈ (lm, lm+1) ; m = 1, . . . , n + 1, the coupling conditions [ αk m1 d dx + βk m1 ] ϕk = [ αk m2 d dx + βk m2 ] ϕk+1, x = lk, k = 1, . . . , n; m = 1, 2, and the boundary conditions ϕ1|x=−∞ = 0 , ϕn+1|x=∞ = 0. Similarly, the eigenfunction ϕ∗ (x, λ), ϕ∗ (ξ, λ) = n∑ k=2 θ (ξ − lk−1) θ (lk − ξ) ϕ∗k (ξ, λ) + θ (l1 − ξ) ϕ∗1 (ξ, λ) + θ (ξ − ln) ϕ∗n+1 (ξ, λ) , is the solution of the system of separated differential equations ( a2 m d2 dx2 + λ2 ) ϕ∗m (x, λ) = 0, x ∈ (lm, lm+1) ; m = 1, . . . , n + 1, the coupling conditions 1 ∆1,k [ αk m1 d dx + βk m1 ] ϕ∗k = 1 ∆2,k [ αk m2 d dx + βk m2 ] ϕ∗k+1, x = lk, where ∆i,k = det ( αk 1i βk 1i αk 2i βk 2i ) k = 1, . . . , n; i,m = 1, 2, and the boundary conditions ϕ1|x=−∞ = 0 , ϕn+1|x=∞ = 0. 596 Journal of Mathematical Physics, Analysis, Geometry, 2013, vol. 9, No. 4 Integral Transforms with Non-separated Variables We normalize the eigenfunctions as follows: ϕn+1 (x, λ) = eia−1 n+1xλ. ϕ∗ n+1 (x, λ) = e−ia−1 n+1xλ. Let Fn and F−1 n be the direct and the inverse Fourier transforms on the Cartesian axis with n division points defined as (see [10]) : Fn [f ] (λ) = n+1∑ m=0 lm∫ lm−1 u∗m (ξ, λ) fm (ξ) dξ ≡ f̂ (λ) , (3) fk (x) = 1 πi ∞∫ 0 uk (x, λ) f̂ (λ) λdλ. (4) 3. The Main Result We will use the method of delta-like functions [4]. This means that we are looking for the solution of the problem, defined by the separated matrix systems (n + 1) of parabolic equations ( ∂ ∂t −A2 j ∂2 ∂x2 −∆y ) Uj (t, x, y) = 0, (t, x, y) ∈ D+ ×Rm, j = 1, n + 1 (5) bounded on the set D ×Rm, D+ = (0,∞)× In, where In = { x : x ∈ n+1 U j=1 (lj−1, lj) , l0 = −∞, ln+1 = ∞, lj < lj+1, j = 1, n } ∆y = ∂2 ∂y2 1 + . . . + ∂2 ∂y2 m , Aj = ( aj kl ) is a positive-definite matrix r × r by the initial conditions Uj (t, x, y) |t=0 = gj (x, y) , x ∈ In, y ∈ Rm, (6) by the edge conditions U1|x=−∞ = 0 , Un+1|x=∞ = 0, (7) and by the coupling conditions [ αk m1 ∂ ∂x + βk m1 ] Uk = [ αk m2 ∂ ∂x + βk m2 ] Uk+1, (8) Journal of Mathematical Physics, Analysis, Geometry, 2013, vol. 9, No. 4 597 O.E. Yaremko x = lk, k = 1, n; m = 1, 2. Here Uj(t, x, y) is an unknown vector-function, gj(x, y) is a given vector- function, αk mi, βk mi, γ k mi, δk mi are the matrices r × r. By using the Fourier integral with discontinuous coefficients of Section 2 and the Fourier integral with non-separated variables (1), (2), we obtain the repre- sentation for the solution of (3)–(6): Uk (t, x, y) = − 1 πi 1(√ 2π )m ∫ Rm n+1∑ j=1 lj∫ lj−1 lim τ→0   ∞∫ 0 Jm−2 2 (λ |y − η|) |y − η| m−2 2 e−λτλ m 2 dλ × ∞∫ −∞ e−β2tϕk (x, β)ϕ∗j (ξ, β) dβ   fj (ξ, η) dξdη, k = 1, n + 1, (9) where ϕk (x, β) , ϕ∗j (ξ, β) are the eigenfunctions of the direct and the dual Sturm –Liouville problems, respectively. We write the integral ∞∫ 0 Jm−2 2 (λ |y − η|) |y − η| m−2 2 e−λτλ m 2 dλ ∞∫ −∞ e−β2tϕk (x, β) ϕ∗j (ξ, β) dβ in the polar coordinates λ = ρ sinϕ, β = ρ cosϕ, 0 ≤ ρ < ∞, 0 ≤ ϕ ≤ π to obtain ∞∫ 0 ρ m 2 ρdρ π∫ 0 e−ρ2t cos2 α sin m 2 α Jm−2 2 (ρ sinα |y − η|) |y − η| m−2 2 e−ρτ sin α ϕk (x, ρ cosα) ϕ∗j (ξ, ρ cosα) dα. We carry out the limit τ → 0 in (9) yielding Uk (t, x, y) = − 1 πi ( 1 2π )m 2 ∫ Rm n+1∑ j=1 lj∫ lj−1 ρ m 2 dρ   π∫ 0 e−ρ2t cos2 α sin m 2 α × Jm−2 2 (ρ sinα |y − η|) |y − η|m−2 2 φk (x, ρ cosα) φ∗j (ξ, ρ cosα) dα ) fj (ξ, η) dξdη. 598 Journal of Mathematical Physics, Analysis, Geometry, 2013, vol. 9, No. 4 Integral Transforms with Non-separated Variables If, in addition, we assume that we can carry out the limit t → 0 in the expan- sion of the eigenfunctions for multidimensional direct Sturm–Liouville problem fk (x, y) of Sec. 2, we obtain fk (x, y) = − 1 πi 1(√ 2π )m ∫ Rm n+1∑ j=1 lj∫ lj−1 ∞∫ 0 ρ m 2 ρdρ   π∫ 0 si n m 2 α Jm−2 2 (ρ sinα |y − η|) |y − η| m−2 2 ×ϕk(x, ρ cosα)ϕ∗j (ξ, ρ cosα)dα ) fj (ξ, η) dξdη. (10) Let us denote ϕk,j ≡ ϕk,j (ρ, x, ξ, |y − η|) = π∫ 0 sin m 2 α Jm−2 2 (ρ sinα |y − η|) |y − η| m−2 2 ×ϕk (x, ρ cosα) ϕ∗j (ξ, ρ cosα) dα. It is clear then that formula (10) can be written as fk (x, y) = 1 π ∞∫ 0 ρ m+1 2 dρ 1(√ 2π )m ∫ Rm n+1∑ j=1 lj∫ lj−1 ϕk,jfj (ξ, η) dξdη. (11) Equation (11) allows us to write down the direct and the inverse multidimensional Fourier transforms with discontinuities on the planes x = lk: Fn [f ] (x, y, λ) = 1(√ 2π )m ∫ Rm n+1∑ j=1 lj∫ lj−1 ϕk,j (λ, x, ξ, |y − η|) fj (ξ, η) dξdη, (12) f(x, y) = ∞∫ 0 λ m 2 +1Fn [f ] (x, y, λ) dλ. (13) Now we can prove the basic integral identity for differential operator B = θ (l1 − xt) ( A2 1 d2 dx2 + ∆y ) + n∑ k=1 θ (x− lk−1) θ (lk − x) ( A2 k d2 dx2 + ∆y ) + θ (x− ln) ( A2 n+1 d2 dx2 + ∆y ) . Journal of Mathematical Physics, Analysis, Geometry, 2013, vol. 9, No. 4 599 O.E. Yaremko Theorem 1. Let f (x, y) = θ (l1 − x) f1 (x, y) + n∑ k=2 θ (x− lk−1) θ (lk − x) fk (x, y) + θ (x− ln) fn+1 (x, y) be a twice continuously differentiable on D+ ×Rm vector-function, in which fn+1(x, y), ∂fn+1(x, y) ∂x vanishes as x → +∞ and y is fixed, f1(x, y), ∂f1(x, y) ∂x vanishes as x → −∞ and y is fixed, fi(x, y), ∂fi(x, y) ∂yj vanishes as yj → ±∞ and x, y1, y2, . . . , yj−1, yj+1, . . . , ym are fixed. Assume also that the coupling conditions (8) are valid. Then the following holds true: Fn [B (f)] = −λ2Fn [f ] . P r o o f. Integrate twice by parts with respect to each of the variables on the left, taking into account the conditions of the theorem. As a result, the operator B acts on the kernel: Fn [B(f)] (x, y, λ) = 1(√ 2π )m ∫ Rm n+1∑ j=1 lj∫ lj−1 Bj [ϕk,j (λ, x, ξ, |y − η|)] fj (ξ, η) dξdη. Let us prove the equality Bj [ϕk,j ] = −λ2ϕk,j . We have Bj [ϕk,j ] = π∫ 0 sin m 2 α∆η  Jm−2 2 (ρ sinα |y − η|) |y − η| m−2 2   ×ϕk (x, ρ cosα) ( ϕ∗j (ξ, ρ cosα) ) dα + π∫ 0 sin m 2 α  Jm−2 2 (ρ sinα |y − η|) |y − η| m−2 2   600 Journal of Mathematical Physics, Analysis, Geometry, 2013, vol. 9, No. 4 Integral Transforms with Non-separated Variables ×ϕk (x, ρ cosα) a2 j ∂2 ∂ξ2 ( ϕ∗j (ξ, ρ cosα) ) dα = −ρ2 sin2 α π∫ 0 sin m 2 α  Jm−2 2 (ρ sinα |y − η|) |y − η| m−2 2   ×ϕk (x, ρ cosα) ( ϕ∗j (ξ, ρ cosα) ) dα −ρ2 cos2 α π∫ 0 sin m 2 α  Jm−2 2 (ρ sinα |y − η|) |y − η| m−2 2   ×ϕk (x, ρ cosα) ( ϕ∗j (ξ, ρ cosα) ) dα = −ρ2ϕk,j . We have used above that ϕ∗j (ξ, ρ cosα) are the eigenfunctions of the dual Sturm– Liouville problems and the relation ∆η  Jm−2 2 (ρ sinα |y − η|) |y − η| m−2 2   = −ρ2 sin2 α  Jm−2 2 (ρ sinα |y − η|) |y − η| m−2 2   . By using the basic identity [9], we conclude ρ m 2 Jm−2 2 (ρ |y|) |y| m−2 2 = 1 (2π) m 2 ∫ Sρ ei<y,ξ>dSρ. This completes the proof. The above formulas for the direct and the inverse Fourier transforms with non-separated variables are significantly simpler in the case of ideal coupling conditions on one surface. This case is widely known in engineering practice. Consider, for the sake of simplicity, the scalar case, assuming that the ideal coupling conditions are in the plane x = 0, ϕ1 (x, y) = ϕ2 (x, y) , x = 0, y ∈ Rm; ϕ′1x (x, y) = νϕ′2x (x, y) , x = 0, y ∈ Rm; ν = λ2 λ1 . The one-dimensional components of eigenfunctions are given in [4]: ϕ1 (x, λ) = ( cosλ x a1 + i 1√ δ0 sinλ x a1 ) (1 + δ0) ; ϕ2 (x, λ) = ( cosλ x a2 + i √ δ0 sinλ x a2 ) (1 + δ0) ; Journal of Mathematical Physics, Analysis, Geometry, 2013, vol. 9, No. 4 601 O.E. Yaremko ϕ∗k (x, λ) = rkϕk (x, λ), k = 1, 2, r1 = a2 ν0a2 1 , r2 = 1 a2 , δ0 = a2 ν0a1 . The multidimensional components of eigenfunctions with non- separated variables ϕkj have the form: ϕ11 = 1 + δ0 a1 Jm−1 2 ( ρ √ (x−ξ)2 a2 1 + |y − η|2 ) ( (x−ξ)2 a2 1 + |y − η|2 )m−1 2 −1− δ0 a1 Jm−1 2 ( ρ √ (x+ξ)2 a2 1 + |y − η|2 ) ( (x+ξ)2 a2 1 + |y − η|2 )m−1 2 , ϕ12 = 1 + δ0 a2 √ δ0 Jm−1 2 ( ρ √( x a2 − ξ a1 )2 + |y − η|2 ) (( x a2 − ξ a1 )2 + |y − η|2 )m−1 2 + 1− δ0 a2 √ δ0 Jm−1 2 ( ρ √( x a2 + ξ a1 )2 + |y − η|2 ) (( x a2 + ξ a1 )2 + |y − η|2 )m−1 2 , ϕ21 = √ δ0 1 + δ0 a1 Jm−1 2 ( ρ √( x a1 − ξ a2 )2 + |y − η|2 ) (( x a1 − ξ a2 )2 + |y − η|2 )m−1 2 + √ δ0 1− δ0 a1 Jm−1 2 ( ρ √( x a1 + ξ a2 )2 + |y − η|2 ) (( x a1 + ξ a2 )2 + |y − η|2 )m−1 2 , ϕ22 = 1 + δ0 a2δ0 Jm−1 2 ( ρ √ (x−ξ)2 a2 2 + |y − η|2 ) ( (x−ξ)2 a2 2 + |y − η|2 )m−1 2 −1− δ0 a2δ0 Jm−1 2 ( ρ √ (x+ξ)2 a2 2 + |y − η|2 ) ( (x+ξ)2 a2 2 + |y − η|2 )m−1 2 . Now the integral transforms given by formulas (12), (13) are constructed. 602 Journal of Mathematical Physics, Analysis, Geometry, 2013, vol. 9, No. 4 Integral Transforms with Non-separated Variables 4. Conclusion Let us remark that integral transforms (12), (13) can be used to solve problems of mathematical physics by using the standard algorithm: find the solution in the images then return to the originals. An advantage of our formulas is that they involve just one spectral parameter contained in the final formulas while the integral transforms with separated variables contain m parameters. References [1] Ja.S. Ufljand, Integral Transforms for the Elasticity Theory Problems. Nauka, Leningrad, 1967. (Russian) [2] Ja.S. Ufljand, Dual Equations Method in of Mathematical Physics Problems. Nauka, Leningrad, 1977. (Russian) [3] M.P. Lenjuk, Integral Hybrid Transforms (Bessel, Legendre). — Ukr. Mat. Zh. 43 (1991), No. 6, 770–779. (Russian) [4] M.P. Lenjuk, Fourier Integral Transform in Piecewise Homogeneous Half-Line. — Izv. vuzov. Matematika 4 (1989), 14–18. (Russian) [5] L.S Najda, Hankel–Legendre Type Integral Hybrid Transforms. — Math. Meth. Dynamic Systems Anal. Kharkov, 8 (1984), 132–135. (Russian) [6] V.S. Procenko and A.I. Solov’jov, Some Integral Hybrid Transforms and their Appli- cations to the Elasticity Theory by Heterogeneous Media. — Prikladnaja mehanika 13 (1982), No. 1, 62–67. (Russian) [7] I.I. Bavrin and O.E. Yaremko, Integral Fourier Transform on Compact Subset of RN and their Applications to the Moment Problem. — Pleiades Publishing, Ltd. Doclady mathematics 374 (2000), No. 2, 177–179. [8] A.F. Nikiforov and V.B. Uvarov, Special Functions of Mathematical Physics: A Uni- fied Introduction with Applications. Publisher: Birkhauser, Boton, 1988. [9] I.N. Sneddon, Fourier Transforms. McGraw-Hill, New York, 1951. [10] O.E. Yaremko, Matrix Integral Fourier Transform for Problems with Discontinuous Coefficients and Transformation Operators. — Pleiades Publishing, Ltd. Doclady mathematics 76 (2007), No. 3, 876–878. Journal of Mathematical Physics, Analysis, Geometry, 2013, vol. 9, No. 4 603