Distributed-order calculus: An operator-theoretic interpretation

Within the Bochner – Phillips functional calculus and the Hirsch functional calculus, we describe the operators of distributed order differentiation and integration as functions of the classical differentiation and integration operators, respectively.

Gespeichert in:
Bibliographische Detailangaben
Datum:2008
1. Verfasser: Kochubei, A.N.
Format: Artikel
Sprache:English
Veröffentlicht: Інститут математики НАН України 2008
Schriftenreihe:Український математичний журнал
Schlagworte:
Online Zugang:https://nasplib.isofts.kiev.ua/handle/123456789/164496
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:Distributed-order calculus: An operator-theoretic interpretation / A.N. Kochubei // Український математичний журнал. — 2008. — Т. 60, № 4. — С. 478–486. — Бібліогр.: 16 назв. — англ.

Institution

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id nasplib_isofts_kiev_ua-123456789-164496
record_format dspace
spelling nasplib_isofts_kiev_ua-123456789-1644962025-02-10T00:45:44Z Distributed-order calculus: An operator-theoretic interpretation Числення розподіленого порядку: теоретико-операторна інтерпретація Kochubei, A.N. Статті Within the Bochner – Phillips functional calculus and the Hirsch functional calculus, we describe the operators of distributed order differentiation and integration as functions of the classical differentiation and integration operators, respectively. У межах функціональних числень Boxнepa - Філліпса та Хірша наведено опис операторів диференціювання та інтегрування розподіленого порядку як функцій від класичних операторів диференціювання та інтегрування. Partially supported by the Ukrainian Foundation for Fundamental Research (Grant 14.1/003). 2008 Article Distributed-order calculus: An operator-theoretic interpretation / A.N. Kochubei // Український математичний журнал. — 2008. — Т. 60, № 4. — С. 478–486. — Бібліогр.: 16 назв. — англ. 1027-3190 https://nasplib.isofts.kiev.ua/handle/123456789/164496 517.9 en Український математичний журнал application/pdf Інститут математики НАН України
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
language English
topic Статті
Статті
spellingShingle Статті
Статті
Kochubei, A.N.
Distributed-order calculus: An operator-theoretic interpretation
Український математичний журнал
description Within the Bochner – Phillips functional calculus and the Hirsch functional calculus, we describe the operators of distributed order differentiation and integration as functions of the classical differentiation and integration operators, respectively.
format Article
author Kochubei, A.N.
author_facet Kochubei, A.N.
author_sort Kochubei, A.N.
title Distributed-order calculus: An operator-theoretic interpretation
title_short Distributed-order calculus: An operator-theoretic interpretation
title_full Distributed-order calculus: An operator-theoretic interpretation
title_fullStr Distributed-order calculus: An operator-theoretic interpretation
title_full_unstemmed Distributed-order calculus: An operator-theoretic interpretation
title_sort distributed-order calculus: an operator-theoretic interpretation
publisher Інститут математики НАН України
publishDate 2008
topic_facet Статті
url https://nasplib.isofts.kiev.ua/handle/123456789/164496
citation_txt Distributed-order calculus: An operator-theoretic interpretation / A.N. Kochubei // Український математичний журнал. — 2008. — Т. 60, № 4. — С. 478–486. — Бібліогр.: 16 назв. — англ.
series Український математичний журнал
work_keys_str_mv AT kochubeian distributedordercalculusanoperatortheoreticinterpretation
AT kochubeian čislennârozpodílenogoporâdkuteoretikooperatornaínterpretacíâ
first_indexed 2025-12-02T07:06:48Z
last_indexed 2025-12-02T07:06:48Z
_version_ 1850379294147084288
fulltext UDC 517.9 A. N. Kochubei (Inst. Math. Nat. Acad. Sci. Ukraine, Kyiv) DISTRIBUTED ORDER CALCULUS: AN OPERATOR-THEORETIC INTERPRETATION ∗∗∗∗ ÇYSLENNQ ROZPODILENOHO PORQDKU: TEORETYKO-OPERATORNA INTERPRETACIQ Within the Bochner – Phillips functional calculus and the Hirsch functional calculus, we describe the operators of distributed order differentiation and integration as functions of the classical differentiation and integration operators, respectively. U meΩax funkcional\nyx çyslen\ Boxnera – Fillipsa ta Xirßa navedeno opys operatoriv dyferencigvannq ta intehruvannq rozpodilenoho porqdku qk funkcij vid klasyçnyx operatoriv dyferencigvannq ta intehruvannq. 1. Introduction and preliminaries. In the distributed order calculus [1], used in phy- sics for modeling ultraslow diffusion and relaxation phenomena, we consider deriva- tives and integrals of distributed order. The definitions are as follows. Let µ be a continuous non-negative function on [0, 1]. The distributed order de- rivative D( )µ of weight µ for a function ϕ on [0, T] is ( )( ) ( )D µ ϕ t = 0 1 ∫ ( )( ) ( ) ( )D α ϕ µ α αt d (1) where D( )α is the Caputo – Dzhrbashyan regularized fractional derivative of order α , that is ( )( ) ( )D α ϕ t = 1 1 0 0 Γ( ) ( ) ( ) ( ) − − −         − −∫α τ ϕ τ τ ϕα αd dt t d t t , 0 < t < T . (2) Denote k ( s ) = s d − −∫ α α µ α α Γ( ) ( ) 1 0 1 , s > 0. (3) It is obvious that k is a positive decreasing function. The definition (1), (2) can be rewritten as ( )( ) ( )D µ ϕ t = d dt k t d k t t ( ) ( ) ( ) ( )− −∫ τ ϕ τ τ ϕ 0 0 . (4) The right-hand side of (4) makes sense for a continuous function ϕ, for which the de- rivative d dt k t d t ( ) ( )−∫ τ ϕ τ τ 0 exists. If a function ϕ is absolutely continuous, then ( )( ) ( )D µ ϕ t = k t d t ( ) ( )− ′∫ τ ϕ τ τ 0 . (5) ∗ Partially supported by the Ukrainian Foundation for Fundamental Research (Grant 14.1/003). © A. N. KOCHUBEI, 2008 478 ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 4 DISTRIBUTED ORDER CALCULUS: AN OPERATOR-THEORETIC INTERPRETATION 479 Below we always assume that µ ∈C3 0 1[ , ], µ ( 1 ) ≠ 0, and either µ ( 0 ) ≠ 0, or µ ( α ) ∼ a α ν, a, ν > 0, as α → 0. Under these assumptions (see [1]), k ( s ) ∼ s s− −1 2 1(log ) ( )µ , s → 0, k ′ ( s ) ∼ – s s− −2 2 1(log ) ( )µ , so that k L T∈ 1 0( , ) and k does not belong to any Lp , p > 1. We cannot differenti- ate under the integral in (4), since k ′ has a non-integrable singularity. It is instructive to give also the asymptotics of the Laplace transform K ( z ) = k s e dszs( ) − ∞ ∫ 0 . Using (4) we find that K ( z ) = z dα µ α α−∫ 1 0 1 ( ) , so that K ( z ) can be extended analytically to an analytic function on C R\ − , R− = = z z z∈ = ≤{ }C : Im , Re0 0 . If z ∈ −C R\ , z → ∞ , then [1] K ( z ) = µ( ) log log 1 2 z O z+ ( )( )− ; (6) see [1] for further properties of K . The distributed order integral I( )µ is defined as the convolution operator ( )( ) ( )I µ f t = κ( ) ( )t s f s ds t −∫ 0 , 0 ≤ t ≤ T, (7) where κ ( t ) is the inverse Laplace transform of the function z z z � 1 K ( ) , κ ( t ) = d dt i e z z z dz zt i i 1 2 1 π γ γ K ( ) − ∞ + ∞ ∫ , γ > 0. (8) It was proved in [1] that κ ∈ ∞∞C ( , )0 and κ is completely monotone; for small values of t, κ ( t ) ≤ C t log1, ′κ ( )t ≤ Ct t −1 1log . (9) If f L T∈ 1 0( , ), then D I ( ) ( )µ µ f = f . The aim of this paper is to clarify the operator-theoretic meaning of the above con- structions. It is well known that fractional derivatives and integrals can be interpreted as fractional powers of the differentiation and integration operators in various Banach spaces; see, for example, [2 – 5]. ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 4 480 A. N. KOCHUBEI Let A be the differential operator Au = – du dx in L Tp( , )0 , 1 ≤ p < ∞ , with the boundary condition u ( 0 ) = 0. Its domain D ( A ) consists of absolutely continuous functions u L Tp∈ ( , )0 , such that u ( 0 ) = 0 and ′ ∈u L Tp( , )0 . We show that on D ( A ) the distributed order differentiation coincides with the function L ( )− A of the operator – A, where L ( z ) = z K ( z ) , and the function of an operator is understood in the sense of the Bochner – Phillips functional calculus (see [6 – 8]). Moreover, if p = 2 then the distributed order integration operator I ( )µ equals N ( J ) , where N ( x ) = 1 L( )x , J is the integration operator, ( Ju ) ( t ) = u d t ( )τ τ 0∫ . This result is obtained within Hirsch’s functional calculus [9, 10] giving more detailed re- sults for a more narrow class of functions. As by-products, we obtain an estimate of the semigroup generated by – L ( )− A , and an expression for the resolvent of the operator I( )µ . 2. Functions of the differentiation operator. The semigroup Ut of operators on the Banach space X = L Tp( , )0 generated by the operator A has the form ( )( )U f xt = f x t t x T x t ( ), , , , − ≤ ≤ < < <    if if 0 0 0 x T∈( , )0 , t ≥ 0. This follows from the easily verified formula for the resolvent R ( λ, A ) = ( )A I− −λ 1 of the operator A : ( R ( λ, A ) u ) ( x ) = – e u y dyx y x − −∫ λ( ) ( ) 0 ; (10) see [11] for a similar reasoning for operators on Lp( , )0 ∞ . The semigroup Ut is nil- potent, Ut = 0 for t > T ; compare Sect. 19.4 in [12]. It follows from the expression (10) and the Young inequality that R A( , )λ ≤ λ−1, λ > 0, so that Ut is a C0-se- migroup of contractions. In the Bochner – Phillips functional calculus, for the operator A, as a generator of a contraction semigroup, and any function f of the form f ( x ) = ( ) ( )1 0 − + +− ∞ ∫ e dt a bxtx σ , a, b ≥ 0, (11) where σ is a measure on ( 0 , ∞ ) , such that t t dt 1 0 + ∞ ∫ σ( ) < ∞ , the subordinate C0-semigroup Ut f is defined by the Bochner integral Ut f = ( ) ( )U u dss tσ 0 ∞ ∫ where the measures σt are defined by their Laplace transforms, ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 4 DISTRIBUTED ORDER CALCULUS: AN OPERATOR-THEORETIC INTERPRETATION 481 e dssx t − ∞ ∫ σ ( ) 0 = e t f x− ( ) . The class B of functions (11) coincides with the class of Bernstein functions, that is functions f C C∈ ∞ ∞∞[ , ) ( , )0 0∩ , for which ′f is completely monotone. Below we show that L ∈ B . The generator A f of the semigroup Ut f is identified with – f A( )− . On the do- main D ( A ) , A uf = – au bAu U u u dtt+ + − ∞ ∫ ( ) ( )σ 0 , u ∈ D ( A ) . (12) Theorem 1. (i) If u ∈ D ( A ) , then A L u = – D ( )µ u . (ii) The semigroup Ut L decays at infinity faster than any exponential function: Ut L ≤ C er rt− for any r > 0. (13) The operator AL has no spectrum. (iii) The resolvent R A( ),λ − L of the operator – AL has the form ( ( ) ), ( )R A u xλ − L = r x s u s ds x λ( ) ( )−∫ 0 , u ∈ X , (14) where r sλ( ) = 1 λ λ d ds u s( ), (15) and uλ is the solution of the Cauchy problem D ( )µ λu = λ λu , uλ( )0 = 1. (16) (iv) The inverse ( )− −AL 1 coincides with the distributed order integration ope- rator I( )µ . (v) The resolvent of I( )µ has the form ( )( ) I µ λ− −I u1 = – 1 1 2 1λ λ λu r u− ∗/ , λ ≠ 0. (17) Proof. Let σ( )dt = – ′k t dt( ) . By (3), ′k t( ) = – α α µ α α αt d − − −∫ 1 0 1 1Γ( ) ( ) , so that t t dt 1 0 + ∞ ∫ σ( ) = αµ α α α σ α( ) ( ) ( ) Γ 1 1 0 1 0 − +∫ ∫ −∞ d t t dt . Using the integral formula 2.2.5.25 from [13] we find that t t dt 1 0 + ∞ ∫ σ( ) = π αµ α απ α α( ) (sin ) ( )Γ 1 0 1 −∫ d < ∞ . ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 4 482 A. N. KOCHUBEI Let us compute the function (11) with a = b = 0. We have f ( x ) = – ( ) ( )1 0 − ′− ∞ ∫ e k t dttx = x e k t dttx− ∞ ∫ ( ) 0 = x K ( x ) = L ( x ) . The corresponding expression (12) for A L u , u ∈ D ( A ) , is as follows: ( A L u ) ( x ) = – [ ]( )( ) ( ) ( )U u x u x k t dtt − ′ ∞ ∫ 0 = = – [ ]( ) ( ) ( ) ( ) ( )u x t u x k t dt u x k t dt x x − − ′ + ′∫ ∫ ∞ 0 = = – k x u x u x t u x k t dt x ( ) ( ) ( ) ( ) ( )[ ]− − − ′∫ 0 . By (4), we find that A L u = – D ( )µ u , u ∈ D ( A ) . The function L ( z ) is holomorphic for Re z > 0. We will need a detailed infor- mation (refining (6)) on the behavior of Re L ( σ + i τ ) , σ , τ ∈ R , σ > 0, when τ → → ∞ . We have Re L ( σ + i τ ) = ϕ α σ τ µ α α( , , ) ( )d 0 1 ∫ where ϕ α σ τ( , , ) = ( ) / cos arctanσ τ α τ σ α2 2 2+     . We check directly that ϕ α σ( , , )0 = σα , ∂ ∂ ϕ α σ τ τ ( , , ) = α σ τ α τ σ τ σ α τ σ α( ) / cos arctan tan arctan2 2 2 1+     −         − ≥ 0, and ∂ ∂ ϕ α σ τ τ ( , , ) > 0 for α < 1. This means that the function gσ τ( ) = Re L ( σ + i τ ) (which is even in τ ) is strictly monotone increasing in τ for τ > 0. Its minimal va- lue is gσ( )0 = σ µ α αα ( )d 0 1 ∫ . On the other hand, Re L ( σ + i τ ) ≥ ( ) / cos ( )σ τ απµ α αα2 2 2 0 1 2 +∫ d = 2 22 2 0 2 π σ τ µ π π π ( ) / / cos+    ∫ t t t dt = = 2 2 0 2 π µ π π e t t dtqt    ∫ cos / = 2 1 22 0 2 π µ π π π e e s s dsq qs/ / sin− −   ∫ where q = 1 2 2 π σ τlog( )+ . By Watson’s asymptotic lemma (see [14]), since µ ( 1 ) ≠ ≠ 0, we have ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 4 DISTRIBUTED ORDER CALCULUS: AN OPERATOR-THEORETIC INTERPRETATION 483 e s s dsqs− −   ∫ µ π π 1 2 0 2 sin / ∼ Cq−2 where C does not depend on σ, τ . Roughening the estimate a little we find that e t i− +Re ( )L σ τ ≤ Ce t− −ρ τ ε1 2/ (18) where 0 < ε < 1 / 2 can be taken arbitrarily, and the positive constants C and ρ do not depend on σ and τ. It follows from (18) (see [15]) that for each t > 0 the function x e t x� − L ( ) is represented by an absolutely convergent Laplace integral. This means that the measure σt ds( ) has a density m ( t, s ) with respect to the Lebesgue measure. Moreover, m ( t, s ) = 1 2π γ γ i e e dzzs t z i i − − ∞ + ∞ ∫ L ( ) , γ > 0. (19) Since Ut = 0 for t > T, we have U ut f = ( ) ( , )U u m t s dss T 0 ∫ . (20) The representation (19) yields the expression m ( t, s ) = e e e d s i t i γ τ γ τ π τ− + ∞ ∫ L ( ) 0 , L ( γ + i τ ) = ( ) ( )γ τ µ α αα+∫ i d 0 1 , 0 ≤ τ < ∞ . We have m t s( , ) ≤ e e d s tgγ τ π τγ− ∞ ∫ ( ) 0 . The above monotonicity property of gγ makes it possible to apply to the last in- tegral the Laplace asymptotic method [14]. We obtain that, for large values of t, m t s( , ) ≤ Ct e es tg− −1 0γ γ ( ) . Changing γ and C we can make the coefficient gγ ( )0 arbitrarily big. By (20), this leads to the estimate (13). Due to (13), the resolvent R ( λ, A L ) = – e U dtt t T −∫ λ L 0 , (21) is an entire function, so that A L has no spectrum. It follows from (21) that R ( λ, – A L ) = e U dtt t T λ L 0 ∫ , and if u ∈ X , Re λ ≤ 0, then ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 4 484 A. N. KOCHUBEI ( R ( λ, – A L ) u ) ( x ) = e dt u x s m t s dst x λ 0 0 ∞ ∫ ∫ −( ) ( , ) = r x s u s ds x λ( ) ( )−∫ 0 where r sλ( ) = e m t s dttλ ( , ) 0 ∞ ∫ . (22) For a fixed ω ∈( , )/1 2 1 , let us deform the contour of integration in (19) from the vertical line to the contour Sγ ω, consisting of the arc Tγ ω, = z z z∈ = ≤{ }C : , argγ ωπ , and two rays Γγ ω, ± = z z z∈ = ± ≥{ }C : arg ,ωπ γ . The contour Sγ ω, is oriented in the direction of growth of arg z. By Jordan’s lemma, m t s( , ) = 1 2π γ ω i e e dzzs t z S −∫ L ( ) , . Under this integral, we may integrate in t, as required in (22). We find that r sλ( ) = 1 2π λ γ ω i e z dz zs S L ( ) , −∫ , s > 0 (23) (for Re λ > 0, γ should be taken big enough). If λ = 0, the right-hand side of (23) coincides with that of (8) (see also the formula (3.4) in [1], and we prove that ( )− −AL 1 = I( )µ . For λ ≠ 0, we rewrite (23) as r sλ( ) = 1 2 1 2π λ λ π λ γ ω γ ω i e z z dz i e dzzs S zs S L L ( ) ( ) , , − −∫ ∫ . (24) For 0 < s < T, we have e dzzs Sγ ω, ∫ = – lim arg R zs z R z e dz →∞ = < < ∫ ωπ π , e dzzs z R z = < < ∫ ωπ πarg ≤ 2R e dRscosϕ ωπ π ϕ∫ ≤ 2 1R eRsπ ω ωπ( ) cos− → 0, as R → ∞ . Thus, the second integral in (24) equals zero, and it remains to compare (24) with the formula (2.15) of [1] giving an integral representation of the function uλ . The formula (17) follows from (15) and the general connection between the resol- vents of an operator and its inverse ([11], Chapter 3, formula (6.18)). The theorem is proved. Note that the expression (17) for the resolvent of a distributed order integration operator is quite similar to the Hille – Tamarkin formula for the resolvent of a fractional integration operator (see [12], Sect. 23.16). In our case, the function uλ is a ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 4 DISTRIBUTED ORDER CALCULUS: AN OPERATOR-THEORETIC INTERPRETATION 485 counterpart of the function z E z� α αλ( ) (for the order α case). However, in our situation no analog of the entire function Eα (the Mittag-Leffler function) has been identified so far. Accordingly, our proof of (17) is different from the reasoning in [12]. 3. Functions of the integration operator. In this section we assume that p = 2. Hirsch’s functional calsulus deals with the class R of functions which are conti- nuous on C \ ( – ∞ , 0 ) , holomorphic on C \ ( – ∞ , 0 ] , transform the upper half-plane into itself, and transform the semi-axis ( 0 , ∞ ) into itself. The class R is a subclass of B. Another important class of functions is the class S of Stieltjes functions f ( z ) = a d z + + ∞ ∫ ρ λ λ ( ) 0 , z ∈ C \ ( – ∞ , 0 ] , where a ≥ 0, ρ is a non-decreasing right-continuous function, such that d t t ρ( ) 1 0 + ∞ ∫ < ∞ . If f is a nonzero function from S, then the function ˜( )f z = 1 1f z( )− also belongs to S. If f ∈ S , then the function H zf ( ) = f z( )−1 belongs to R . It has the form H zf ( ) = a z z d+ + ∞ ∫1 0 λ ρ λ( ), z ∈ C \ ( – ∞ , 0 ] . For some classes of linear operators V, the function H Vf ( ) is defined as a closure of the operator Wx = ax V I V x d+ + − ∞ ∫ ( ) ( )λ ρ λ1 0 , x ∈ D ( V ) . In particular, this definition makes sense if – V is a generator of a contraction C0- semigroup, and in this case the above construction is equivalent to the Bochner – Phil- lips functional calculus [3, 16]. In addition, by Theorem 2 of [9], if ( )− −V 1 is also a generator of a contraction C0-semigroup, then [ ]( )H Vf −1 = H V f̃ ( )−1 . (25) In order to apply the above theory to our situation, note that [9] z α = 1 1 1 0 Γ Γ( ) ( )α α λ λ λα − + ∞ −∫ z z d , 0 < α < 1, whence L ( z ) = z z d 1 0 + ∞ ∫ λ β λ λ( ) where β ( λ ) = λ µ α α α α α− −∫ ( ) ( ) ( )Γ Γ 1 0 1 d . ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 4 486 A. N. KOCHUBEI Thus L ( z ) = H zf ( ), with f ( z ) = 1 0 z d + ∞ ∫ λ β λ λ( ) . It follows from Watson’s lemma [14] that β ( λ ) ≤ C (log )λ −2 for large values of λ . Therefore β λ λ λ( ) 1 0 + ∞ ∫ d < ∞ . Denote N ( z ) = H f z˜ ( ) = 1 L ( )z . If V = – A, then ( )− −V 1 = – J, where J is the integration operator. It is easy to check that 〈 + 〉∗( ) ,J J u u ≥ 0 ( 〈⋅ ⋅〉, is the inner product in L T2 0( , )) . Therefore – J is a generator of a contraction semigroup. After these preparations, the equality (25) implies the following result. Theorem 2. The operator I ( )µ of distributed order integration and the integra- tion operator J are connected by the relation I ( )µ = N ( J ) . 1. Kochubei A. N. Destributed order calculus and equations of ultraslow diffusion // J. Math. Anal. and Appl. – 2008. – 340. – P. 252 – 281. 2. Bakaev N. Yu., Tarasov R. P. Semigroups and a method for stably solving the Abel equation // Sib. Math. J. – 1978. – 19. – P. 1 – 5. 3. Gohberg I. C., Krein M. G. Theory and applications of Volterra operators in Hilbert space. – Providence: Amer. Math. Soc., 1970. – X + 430 p. 4. Jacob N., Krägeloh A. M. The Caputo derivative, Feller semigroups, and the fractional power of the first order derivative on C∞ +( )R0 // Fract. Calc. Appl. Anal. – 2002. – 5. – P. 395 – 410. 5. Samko S. G., Kilbas A. A., Marichev O. I. Fractional integrals and derivatives: theory and applications. – New York: Gordon and Breach, 1993. – XXXVI + 976 p. 6. Phillips R. S. On the generation of semigroups of linear operators // Pacif. J. Math. – 1952. – 2. – P. 343 – 369. 7. Berg C., Boyadzhiev Kh., deLaubenfels R. Generation of generators of holomorphic semigroups // J. Austral. Math. Soc. A. – 1993. – 55. – P. 246 – 269. 8. Schilling R. L. Subordination in the sense of Bochner and a related functional calculus // Ibid. – 1998. – 64. – P. 368 – 396. 9. Hirsch F. Intégrales de résolvantes et calcul symbolique // Ann. Inst. Fourier. – 1972. – 22, # 4. – P. 239 – 264. 10. Hirsch F. Domaines d’opérateurs representés comme intégrales de résolvantes // J. Funct. Anal. – 1976. – 23. – P. 199 – 217. 11. Kato T. Perturbation theory for linear operators. – Berlin: Springer, 1966. – XX + 592 p. 12. Hille E., Phillips R. S. Functional analysis and semigroups. – Providence: Amer. Math. Soc., 1957. – XII + 808 p. 13. Prudnikov A. P., Brychkov Yu. A., Marichev O. I. Integrals and series. Vol. 1: Elementary functions. – New York: Gordon and Breach, 1986. – 798 p. 14. Olver F. W. J. Asymptotics and special functions. – New York: Acad. Press, 1974. – XVI + 572 p. 15. Ditkin V. A., Prudnikov A. P. Integral transforms and operational calculus. – Oxford: Pergamon Press, 1965. – XI + 529 p. 16. Gorbachuk V. I., Knyazyuk A. V. Boundary values of solutions of differential operator equations // Rus. Math. Surv. – 1989. – 44, # 3. – P. 67 – 111. Received 19.10.07 ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 4