Abstract
In this research article, we formulate and prove the multidimensional Widder–Arendt theorem and the integrated form of the multidimensional Widder–Arendt theorem for functions with values in sequentially complete locally convex spaces. Established results seem to be new even for scalar-valued functions.
Keywords:
multidimensional vector-valued Laplace transform; multidimensional Widder–Arendt theorem; sequentially complete locally convex spaces MSC:
44A10; 44A30; 30E20; 47D99
1. Introduction and Preliminaries
In 1785, Pierre-Simon, Marquis de Laplace, introduced a new type of integral transform for applications to probability. Speaking matter-of-factly, he sometimes introduced what is nowadays called the transform, and he was the predecessor of what is today called Mellin’s transform. The common opinion is that P.-S. Laplace was not the first to invent his transform. The generating functions, used by Laplace, have already been known from the work of A. de Moivre (1718), who can be called the father of modern harmonic analysis. Furthermore, many historians of mathematics highlight that the modern form of the Laplace transform was probably used for the first time by L. Euler in 1744. The Laplace transform has been systematically investigated by the British physicist Oliver Heaviside (1850–1925).
The Widder representation theorem for the Laplace transform was proved in 1934, [1,2]. Following the ideas of I. Miyadera ([3]), S. Zaidman proved in 1960 that the Widder representation theorem holds for functions with values in a reflexive Banach space X ([4]). In 1987, W. Arendt proved that the Widder representation theorem holds for functions with values in a Banach space X if and only if X has the Radon–Nikodym property, which was essentially important for a further expansion of the theory of abstract ill-posed Cauchy problems ([5]). In 1991, M. Hieber proved an integrated version of the Widder theorem in an arbitrary Banach space ([6]). After that, the Widder–Arendt theorem has been seriously reconsidered by J. Kisyński [7], Y.-C. Li, S.-Y. Shaw [8] and T.-J. Xiao, J. Liang [9]; cf. also [10] (Theorem 2.2.1, Theorem 2.4.1, Theorem 2.4.2). The main aim of this paper is to reconsider the Widder–Arendt theorem for multidimensional vector-valued Laplace transform ([11]).
The structure and main ideas of the paper can be briefly summarized as follows. We first explain the basic notation used throughout the paper and provide some useful observations about the Widder–Arendt theorem in the one-dimensional setting; cf. Theorems 1 and 2 in Section 1.1. We recall the basic definitions and results about the multidimensional vector-valued Laplace transform in Section 1.2. The multidimensional Widder–Arendt theorem in locally convex spaces with Lipschitz–Radon–Nikodym property, which seems to be introduced here for the first time, is formulated and proved in Section 2; cf. Theorem 3. Motivated by results established in [6,8,9], in Section 3 we prove integrated form of the multidimensional Widder–Arendt theorem in an arbitrary sequentially complete locally convex space; cf. Theorem 4. We can freely say that the results established in this paper, like the famous Hille-Yosida theorem, are of pure theoretical interest in functional analysis.
Notation and preliminaries.
By X we denote a Hausdorff sequentially complete locally convex space over the field of complex numbers (SCLCS). The abbreviation ⊛ stands for the fundamental system of seminorms which defines the topology of X; by we denote the dual space of X equipped with the strong topology. For further information concerning the integration of functions with values in SCLCSs, we refer the reader to [12]. If , then we define and
The Gamma function is denoted by and the principal branch is always used to take the powers. Define and ( ); is the Dirac distribution. Generally, the finite convolution product is defined by If then we say that a function is Lipschitz continuous if for each seminorm there exists a real constant such that for all a function is said to be Lipschitz continuous if for each the function is Lipschitz continuous.
The Radon–Nikodym property in locally convex spaces has been investigated by many authors; see, e.g., [13,14,15,16,17] and the doctoral dissertation of Y. J. Yaseen [18]. We say that X has the Lipschitz–Radon–Nikodym property if any Lipschitz continuous function is differentiable almost everywhere, and its first derivative is locally integrable (in the sense of [12] (Definition 1.1.2(i))).
In the setting of complex Banach spaces, the Lipschitz–Radon–Nikodym property is equivalent to the Radon–Nikodym property and the Gel’fand property; it is completely without scope of this paper to further relate these concepts in the setting of locally convex spaces (cf. [10] (Section 1.2), [18] (Definition 3.1, p. 36), [19] and [20] (p. 371) for more details). We will only emphasize here that a Fréchet space X has the Lipschitz–Radon–Nikodym property if any Lipschitz continuous function is differentiable almost everywhere because its first derivative is automatically locally integrable (the assumption for all implies for a.e. and this is sufficient to ensure the local integrability of ); to the best knowledge of the author, this cannot be so easily done in general SCLCSs. Since any Lipschitz continuous function is locally of bounded variation, we consequently obtain that any Fréchet space with the Gel’fand property has the Lipschitz–Radon–Nikodym property ([20]).
1.1. Widder–Arendt Theorem in Locally Convex Spaces
In this subsection, we recall and provide some new results about the Widder–Arendt theorem in locally convex spaces. First of all, we will slightly reformulate the integrated version of Widder–Arendt theorem in locally convex spaces (cf. [9] (Theorem 2.3); part (iii)’ is new and not considered elsewhere):
Theorem 1.
Suppose that , X is an SCLCS, and is a family of positive real numbers. Then the following statements are equivalent:
- (i)
- is infinitely differentiable, and for each seminorm we have
- (i)’
- There exists such that is infinitely differentiable and for each seminorm the estimate (1) holds for any and
- (ii)
- admits an analytical extension to the right half-plane denoted by the same symbol, and for each seminorm we have
- (iii)
- For every , there exists a continuous function such thatfor each seminorm we haveand the function satisfies that for each seminorm we have
- (iii)’
- For every , there exists a continuous function such thatfor each seminorm we haveand the function satisfies that for each seminorm we have
Proof.
The implication (iii) ⇒ (ii) follows from the proof of implication (ii) ⇒ (i) in the above-mentioned theorem, with , the implication (ii) ⇒ (i)’ is trivial, and the implication (i) ⇒ (iii) follows from the proof of implication (i)’ ⇒ (ii) in the above-mentioned theorem, with . Therefore, the statements (i)’, (ii) and (iii) are equivalent. The equivalence of these conditions with (i) follows from the trivial implications (ii) ⇒ (i) ⇒ (i)’. If (iii) holds, then we define the function by Then, due to (i), it follows that for each seminorm we have , Applying (i) ⇒ (iii) to function , with , we get (iii)’. Finally, if (iii)’ holds, then (iii) holds for function defined above, with Hence, (i) holds for function , with , so that for each seminorm we have , This simply implies (i) and completes the proof. □
The following result is not well formulated in the setting of locally convex spaces by now (cf. [10] (Theorem 2.4.2) for the Banach space setting):
Theorem 2.
Suppose that , X is an SCLCS, X has the Lipschitz–Radon–Nikodym property, and is a family of positive real numbers. Then the following statements are equivalent:
- (i)
- (ii)
- There exists such thatand for each seminorm we have .
Proof.
Let us assume first (ii). Then is analytic in the right half-plane and
cf. the equation [12] ((34), p. 58). Then a simple calculation yields that (i) holds. To prove the converse statement, let us consider the function given in part (iii)’ of Theorem 1. Then we know that (2) holds with as well as that for each seminorm the estimate (2) holds with the function replaced by the function therein. Therefore, for each seminorm we have and the function is Lipschitz continuous. Since X has the Lipschitz–Radon–Nikodym property, its first derivative exists for a.e. and it is locally integrable on the non-negative real axis. It is clear that for each seminorm we have for a.e. Applying the formula for partial integration established in [12] (Theorem 1.1.4(iii)) and keeping in mind the equality it readily follows that
This yields
which immediately implies (i). □
1.2. Multidimensional Vector-Valued Laplace Transform
The double Laplace transform of scalar-valued functions was initially analyzed by D. L. Bernstein [21] (1939), J. C. Jaeger [22] (1939–1941) and L. Amerio [23] (1940). In our recent paper [11], we have analyzed the multidimensional Laplace transform of functions with values in SCLCSs. We recall the following notion:
Definition 1.
Suppose that X is an SCLCS, is a locally integrable function and . If
exists in topology of then we say that the Laplace integral exists. We define the region of convergence of the Laplace integral by Furthermore, if for each seminorm we have
then we say that the Laplace integral converges absolutely. We define the region of absolute convergence of the Laplace integral by Finally, we define as the set of all tuples such that the set
is bounded in
If is locally integrable, is open and then the mapping is holomorphic and we have
for all and for the basic information concerning holomorphic vector-valued functions of several variables, we refer the reader to the research article [24] by K. Kruse and a list of references quoted in [11].
We will employ the following result from [11]:
Lemma 1.
Suppose that is a locally integrable function, and
Then, the following Post–Widder inversion formula holds as:
Suppose, finally, that is a locally integrable function, is continuous, and there exist real constants and such that, for every seminorm , there exists a finite real constant such that
for a.e. Then, the following multidimensional Post–Widder inversion formula holds:
2. Multidimensional Widder–Arendt Theorem in Locally Convex Spaces with Lipschitz–Radon–Nikodym Property
The main result of this section reads as follows:
Theorem 3.
Let , X is an SCLCS, X has the Lipschitz–Radon–Nikodym property, is a family of positive real numbers, and Let us consider the following statements:
- (i)
- is infinitely differentiable, and for each seminorm we havefor any and
- (ii)
- There exists such thatand for each seminorm we have .
Then we have (ii) ⇒ (i) . Moreover, if X is a Fréchet space, then (i) ⇔ (ii).
Proof.
If (ii) holds, then the mapping is analytic and (3) holds for and Since for each seminorm we have , (i) follows from a simple computation involving the Fubini theorem.
Assume now that (i) holds and X is a Fréchet space. We will prove that (ii) holds by induction on By Theorem 2, (ii) is true in dimension suppose that (ii) is true in any dimension with the meaning clear. Let us fix and consider the function defined by Since for each seminorm we have
an application of Theorem 2 yields the existence of a locally integrable function such that
and for each seminorm we have
Since X is a Fréchet space, the Post–Widder inversion formula and (5) and (6) together imply that there exists a Lebesgue measurable set whose Lebesgue measure is equal to zero, such that
Now we will prove that, for every the mapping is infinitely differentiable as well as that
for any Suppose first that and Let be such that and let If , then we have
as Repeating this argumentation, we obtain that (7) holds for and
Further on, by the induction hypothesis, for every there exists a locally integrable function such that
and for each seminorm we have
Define now for a.e. ; then we have Using the Fubini theorem, (5) and (9), we get
for Keeping in mind the estimate (9), this implies that (ii) is true in dimension which completes the proof of the theorem. □
3. Integrated Form of Multidimensional Widder–Arendt Theorem in Arbitrary Locally Convex Spaces
In this section, we will prove the following integrated form of Widder–Arendt theorem in arbitrary locally convex spaces:
Theorem 4.
Suppose that , and is a family of positive real numbers. Then the following statements are equivalent:
- (i)
- (i)’
- There exist real numbers such that the function is infinitely differentiable and for each seminorm the estimate (4) holds for any and
- (ii)
- admits an analytical extension to the region denoted by the same symbol, and for each seminorm we havefor any and
- (iii)
- We have (iii.1) and (iii.2) , where:
- (iii.1)
- For every there exists a continuous function such that , for each seminorm we havefor any andfor .
- (iii.2)
- For each seminorm and for each functional , there exists a locally integrable function such that , andfor any .
If this is the case, then we havefor any and - (iii)’
- We have (iii.1)’ and (iii.2)’, where:
- (iii.1)’
- For every there exists a continuous function such that , for each seminorm we haveandfor any .
- (iii.2)’
- For each seminorm and for each functional , there exists a locally integrable function such that , andfor any .
Proof.
It is clear that (ii) implies (i) and (i)’. If (i)’ holds, then we can use the usual series expansion to prove (ii); see also the proof of [8] (Theorem 2.1, p. 78). Therefore, the assertions (i), (i)’ and (ii) are equivalent.
If (iii) holds, then we can use the convolution theorem for Laplace transform and the uniqueness theorem for Laplace transform in order to see that (13) holds for any and cf. [11] for more details. Now we will prove that (iii) implies (ii). Let and . Using (10) and (11), it readily follows that admits an analytical extension to the region Keeping in mind this fact and the uniqueness theorem for analytic functions, it follows that (12) holds for any . Now, by the generalization of the norm formula given in [25] (Proposition 22.14), we have
for any and This yields (ii).
The equivalence between (iii) and (iii)’ can be shown using the same arguments as in the corresponding part of proof of Theorem 1. Therefore, it remains to be proved that (ii) implies (iii). Since for each seminorm we have an application of Theorem 3 implies that (iii.2) holds.
The part (iii.1) can be deduced similarly to the one-dimensional case, and we will only provide the main details of the proof here. Let us consider the space consisting of all linear functionals on such that for each equicontinuous family the fundamental system of seminorms which defines a sequentially complete locally convex topology on is given by where denotes the collection of all equicontinuous families We know that X and are linearly and topologically isomorphic ([9]).
Suppose now that the numbers , are given. Using Theorem 3, it follows that for each functional there exists a function such that
Define
for any and Using the convolution theorem for multidimensional Laplace transform [11] (Proposition 2.17), it follows that
for For every we define by Arguing as in the proof of [9] (Theorem 2.3), we can show that for all as well as that the mapping is continuous with respect to the topology of let us only mention here that the linearity of mapping for fixed follows from the continuity of function for each and the uniqueness theorem for multidimensional Laplace transform [11] (Theorem 2.23). The existence of a continuous function such that and (11) holds for now follows by identification of X and through evaluation and the multidimensional Post–Widder formula given in Lemma 1; cf. also the paragraph following Lemma 1 and [9] (l.-1–l.-3, p. 457). Using the generalized norm formula, we can simply prove that for each seminorm the estimate (10) holds for any □
Remark 1.
Alternatively, the proof of implication (ii) ⇒ (iii.1) can be given following the argumentation proposed by Y.-C. Li and S.-Y. Shaw in the proof of [8] (Theorem 2.1, pp. 79–80). In actual fact, we can use the complex characterization theorem for multidimensional vector-valued Laplace transform [11] (Theorem 2.13) in order to see that, for every real numbers there exists a continuous function such that and
for The function constructed in (iii.1) and the function satisfy the following equality
so that the only problem in this approach is to prove that the partial derivative continuously exists on It is our strong belief that this problem can be solved with the help of estimate (4) and the bipolar theorem.
In the proof of [8] (Theorem 2.1), the authors do not use the identification of pivot space X and a closed subspace of its bidual , which has been originally used in the Banach space setting by M. Hieber in the proof of [6] (Theorem 3.2). The proof of [9] (Theorem 2.3) is also meaningful because T.-J. Xiao and J. Liang have explicitly constructed the space which is linearly and topologically isomorphic to the initial space X in the locally convex space setting.
In [26], we have recently introduced the class of multidimensional -regularized C-resolvent solution operator families subgenerated by multivalued linear operators. The class of exponentially equicontinuous -regularized C-resolvent solution operator families can be precisely characterized with the help of a multidimensional vector-valued Laplace transform. The multidimensional Widder–Arendt type theorems established in this paper can serve to provide the necessary and sufficient conditions for the generation of multidimensional -regularized C-resolvent solution operator families. These conditions can be clarified in a rather technical way; cf. also [12] (Theorem 3.2.12) for the one-dimensional setting.
4. Conclusions and Final Remarks
In this paper, we have proved a multidimensional Widder–Arendt theorem and an integrated form of the multidimensional Widder–Arendt theorem for functions with values in sequentially complete locally convex spaces. We have specifically considered the class of locally convex spaces with Lipschitz–Radon–Nikodym property. The results obtained in this paper seem to be new even for scalar-valued functions.
Finally, let us observe that it would also be interesting to extend the statement of the Widder–Arendt–Kisýnski theorem ([7]) and the remarkable results established by A. Bobrowski [27,28] and W. Chojnacki [29,30] to the multidimensional setting, for functions with values in locally convex spaces.
Funding
This research is partially funded by grant 451-03-68/2020/14/200156 of the Ministry of Science and Technological Development, Republic of Serbia.
Data Availability Statement
No data was used for the research described in the article.
Conflicts of Interest
The author declares no conflicts of interest.
References
- Widder, D.V. The inversion of the Laplace integral and the related moment problem. Trans. Am. Math. Soc. 1934, 36, 107–200. [Google Scholar] [CrossRef]
- Widder, D.V. The Laplace Transform; Princeton University Press: Princeton, NJ, USA, 1946. [Google Scholar]
- Miyadera, I. On the representation theorem by the Laplace-transformation of vector-valued functions. Tohoku Math. J. 1956, 8, 170–180. [Google Scholar] [CrossRef] [Scilit]
- Zaidman, S. Sur un théorème de I. Miyadera concernant la représentation des fonctions vectorielles par des intégrales de Laplace. Tohoku Math. J. 1960, 12, 47–51. [Google Scholar] [CrossRef] [Scilit]
- Arendt, W. Vector-valued Laplace transforms and Cauchy problems. Israel J. Math. 1987, 59, 327–352. [Google Scholar] [CrossRef] [Scilit]
- Hieber, M. Laplace transform and α-times integrated semigroups. Forum Math. 1991, 3, 595–612. [Google Scholar] [CrossRef] [Scilit]
- Kisyński, J. Around Widder’s characterization of the Laplace transform of an element of L∞(R+). Ann. Polon. Math. 2000, 74, 161–200. [Google Scholar] [CrossRef] [Scilit]
- Li, Y.-C.; Shaw, S.-Y. N-times integrated C-semigroups and the abstract Cauchy problem. Taiwan. J. Math. 1997, 1, 75–102. [Google Scholar] [CrossRef] [Scilit]
- Xiao, T.-J.; Liang, J. Laplace transforms and integrated, regularized semigroups in locally convex spaces. J. Funct. Anal. 1997, 148, 448–479. [Google Scholar] [CrossRef] [Scilit]
- Arendt, W.; Batty, C.J.K.; Hieber, M.; Neubrander, F. Vector-Valued Laplace Transforms and Cauchy Problems; Monographs in Mathematics; Birkhäuser: Basel, Switzerland, 2001; Volume 96. [Google Scholar]
- Kostić, M. Multidimensional vector-valued Laplace transform and applications. arXiv 2025, arXiv:2506.18909. [Google Scholar]
- Kostić, M. Abstract Degenerate Volterra Integro-Differential Equations; Mathematical Institute SANU: Belgrade, Serbia, 2020. [Google Scholar]
- Blondia, C. On the Radon–Nikodym Property in locally convex spaces and the completeness of . Rev. Real Acad. Cienc. Exactas Fíís. Nat. 1987, 81, 635–647. [Google Scholar]
- Chi, G.Y.H. A geometric characterization of Fréchet spaces with the Radon–Nikodym property. Proc. Am. Math. Soc. 1975, 48, 371–380. [Google Scholar]
- Gilliam, D. On integration and Radon–Nikodym theorem in quasi-complete locally convex topological vector spaces. J. Reine Angew. Math. 1977, 292, 125–137. [Google Scholar]
- Khurana, S.S. Radon–Nikodym property. Comment. Math. Univ. Carol. 2017, 58, 461–464. [Google Scholar]
- Sambucini, A.R. A Radon–Nikodym theorem in locally convex spaces with respect to integration by seminorms. Riv. Mat. Univ. Parma 1995, 5, 49–60. (In Italian) [Google Scholar]
- Yaseen, Y.J. Radon–Nikodym Property in Locally Convex Spaces. Ph.D. Thesis, Department of Mathematics, Royal Holloway and Bedford New College, University of London, London, UK, 1985. [Google Scholar]
- Diestel, J.; Uhl, J.J. Vector Measures; Math. Surveys, No. 15; American Mathematical Society: Providence, RI, USA, 1977. [Google Scholar]
- Grosse-Erdmann, K.-G. Lebesgue’s theorem of differentiation in Fréchet lattices. Proc. Am. Math. Soc. 1991, 112, 371–379. [Google Scholar] [CrossRef] [Scilit]
- Bernstein, D.L. The Double Laplace Integral. Ph.D. Thesis, Brown University, Providence, RI, USA, 1939. [Google Scholar]
- Jaeger, J.C. The solution of boundary value problems by a double Laplace transformation. Bull. Am. Math. Soc. 1940, 46, 687–693. [Google Scholar] [CrossRef] [Scilit]
- Amerio, L. Sulla trasformata doppia di Laplace. Atti Reale Accad. Italia Mem. Cl. Sci. Fis. Mat. Nat. 1941, 7, 707–780. [Google Scholar]
- Kruse, K. Vector-valued holomorphic functions in several variables. Funct. Approx. Comment. Math. 2020, 63, 247–275. [Google Scholar] [CrossRef] [Scilit]
- Meise, R.; Vogt, D. Introduction to Functional Analysis; Ramanujan, M.S., Translator; Revised by the authors; Oxford Graduate Texts in Mathematics; Clarendon Press: New York, NY, USA, 1997. [Google Scholar]
- Kostić, M. Multidimensional (a,k)-regularized C-resolvent solution operator families and applications. Chelj. Phy. Math. J. 2025. Available online: https://www.researchgate.net/publication/391773389 (accessed on 25 May 2025).
- Bobrowski, A. On the Yosida approximation and the Widder–Arendt representation theorem. Stud. Math. 1997, 124, 281–290. [Google Scholar] [CrossRef] [Scilit]
- Bobrowski, A. Inversion of the Laplace transform and generation of Abel summable semigroups. J. Funct. Anal. 2001, 186, 1–24. [Google Scholar] [CrossRef] [Scilit]
- Chojnacki, W. On the equivalence of a theorem of Kisyński and the Hille–Yosida generation theorem. Proc. Am. Math. Soc. 1998, 126, 491–497. [Google Scholar] [CrossRef] [Scilit]
- Chojnacki, W. A generalization of the Widder–Arendt Theorem. Proc. Edinb. Math. Soc. 2002, 45, 161–179. [Google Scholar] [CrossRef] [Scilit]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2026 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license.