Abstract
The asymptotic expansions of the Wright functions of the second kind, introduced by Mainardi [see Appendix F of his book Fractional Calculus and Waves in Linear Viscoelasticity (2010)],
for x → ± ∞ are presented. The situation corresponding to the limit σ → 1− is considered, where Mσ(x) approaches the Dirac delta function δ(x − 1). Numerical results are given to demonstrate the accuracy of the expansions derived in the paper, together with graphical illustrations that reveal the transition to a Dirac delta function as σ → 1−.
1 Introduction
The particular Wright function under consideration (also known as a generalized Bessel function) is defined by
where λ is supposed real and μ is, in general, an arbitrary complex parameter. The series converges for all finite z provided λ > −1 and, when λ = 1, it reduces to the modified Bessel function
The case corresponding to λ = −σ, 0 < σ < 1 arises in the analysis of time-fractional diffusion and diffusion-wave equations. The function with negative λ has been termed a Wright function of the second kind by Mainardi [3], with the function with λ > 0 being referred to as a Wright function of the first kind. In the former context, Mainardi [4, Appendix F] defined the auxiliary functions
These functions are interrelated by the following relation:
The case μ = 0 in (1.1) also finds application in probability theory and is discussed extensively in [13], where it is denoted by
and referred to as a ’reduced’ Wright function.
Plots of Mσ(x) for real x and varying σ are presented in [4, Appendix F] and [5]. These graphs illustrate the transition between the special values
where Ai is the Airy function. As σ → 1−, the function Mσ(x) tends to the Dirac delta function δ(x − 1).
In this paper we present the asymptotic expansions of Fσ(x) and Mσ(x) for x → ±∞ by exploiting the known asymptotics of the function ϕ(−σ, 0, x) discussed in [13]. The resulting expansions involve a combination of algebraic-type and exponential-type expansions, for which explicit representation of the coefficients in both types of expansion is given. In order to give a self-contained account, we describe the derivation of the expansion for Mσ(x) based on the asymptotics of integral functions of hypergeometric type described in [10] (see also [11, §4.2]). The asymptotic treatment of the function Wλ, μ(z) given by Wright [14], [15] did not give precise information about the coefficients appearing in the exponential expansions; see also [10] for a more detailed account.
2 The asymptotic expansions of Fσ(x) and Mσ(x) for x → ±∞
We define the quantities
The connection between Fσ(x) and the function ϕ defined in (1.6) is
The asymptotic expansions of ϕ(−σ, 0, x) for x → ±∞ when 0 < σ < 1 are given in [13, §5.2]. We therefore obtain the expansions stated in the following theorem:
Theorem 2.1
When 0 < σ < 1 we have the expansion of the auxiliary Wright functionFσ(x) given by[*]
and
asx → +∞, whereA′(σ) = A(σ)(σ/κ)κandc0(σ) = 1. The formal exponential and algebraic expansionsE′(x) andH′(x) are defined by (see [13, (5.10), (5.11)])
and
The case
but see the comment at the end of Section 3 as this case is associated with a Stokes phenomenon.
The coefficients cj(σ) appearing in the exponential expansions in Theorem 2.1 can be obtained [**] from [10, (4.6)] (when the parameter δ therein is replaced by σ). We have:
where the first few coefficients dj(σ) are:
These polynomial coefficients are related to the so-called Zolotarev polynomials, see [13].
From the relation (1.5), we have Mσ(±x) = Fσ(±x)/(±πx) and after a little algebra we deduce the expansion of Mσ(x) given by:
Theorem 2.2
When 0 < σ < 1 we have the expansion of the auxiliary Wright functionMσ(x) given by:
and
asx → +∞, where the coefficientscj(σ) are as defined in Theorem 2.1. The formal exponential and algebraic expansions
and
For x → +∞, the function Mσ(x) is exponentially small for all values of σ in the interval 0 < σ < 1. The case of Mσ(−x), however, is seen to be more structured. When
3 The asymptotic expansion of Mσ(x) for x → ±∞
In order to make this paper more self contained we present in this section an alternative derivation of the expansion of Mσ(x) as x → ±∞. Define the function
Then use of the reflection formula for the gamma function shows that the auxiliary Wright function Mσ(x) defined in (1.4) can be expressed in terms of
and in a similar manner
From the discussion in [10, Section 2], the Stokes lines for
From [10, §4.1] (see also [12, §2.3]), the asymptotic expansion of
as |z| → ∞. The upper or lower signs are chosen according as arg z > 0 or arg z < 0, respectively and ϵ denotes an arbitrarily small positive quantity. The formal exponential and algebraic expansions E(z) and H(z) are defined by
where the parameters κ, h, ϑ and A(σ) are defined in (2.1) and the coefficients cj(σ) are those appearing in Theorem 1; see Appendix A for an algorithm for the calculation of these coefficients.
The exponential expansion E(z) is dominant in the sector
3.1 The expansion of Mσ(x) as x → +∞
To deal with this case we require the expansion of
as x → +∞, where Aj(σ) = A(σ)cj(σ) and m denotes the optimal truncation index (that is, truncation at, or near, the smallest term) of the algebraic expansion; see also [9, §4.2]. The coefficients Bj(σ) involve linear combinations of the Aj(σ); see [10, §4.1]. However, the precise values of m and Bj(σ) do not concern us here since in the combination (3.2) the algebraic expansion and the terms involving Bj(σ) all cancel.
The algebraic component of the right-hand side of (3.2) is then seen to be
upon recalling that
3.2 The expansion of Mσ(−x) as x → +∞ (when σ ≠ 1 2 )
The algebraic component in the expansion for Mσ(−x) is from (3.6) and (3.3)
Note that
provided
Remark 3.1
The expansion (2.6) in Theorem 2.2 does not hold when
4 Numerical results
We present some numerical results to verify the expansions in Theorems 1 and 2. In Table 1 the values (accurate to 10dp) of the coefficients cj(σ) appearing in the exponential expansion are shown for two values of σ. Table 2 shows the absolute relative error in the computation of Mσ(x) as a function of the truncation index j with the expansion (2.5) in Theorem 2.2. Table 3 shows the same error in the computation of Mσ(−x) for different values of x with the expansion (2.6). Note that for σ = 1/4 and σ = 1/3 in Table 3 we have
The limit σ → 1− in Mσ(x) can be obtained by setting σ = 1 − ϵ, ϵ → 0+ so that the parameters in (2.1) become
j | σ = 1/4 | σ = 3/4 |
---|---|---|
0 | +1.0000000000 | +1.0000000000 |
1 | +0.1458333333 | −0.0347222222 |
2 | +0.0835503472 | −0.0167582948 |
3 | +0.0597617067 | −0.0224719333 |
4 | +0.0052249186 | −0.0510817883 |
5 | −0.2249669579 | −0.1651975373 |
6 | −1.1657705000 | −0.6952815250 |
σ = 1/4 | σ = 3/4 | |||
---|---|---|---|---|
x = 6 | x = 10 | x = 4 | x = 6 | |
0 | 2.623 × 10−2 | 1.376 × 10−2 | 1.262 × 10−3 | 2.531 × 10−4 |
1 | 2.819 × 10−3 | 7.618 × 10−4 | 2.190 × 10−5 | 8.881 × 10−7 |
2 | 4.123 × 10−4 | 5.561 × 10−5 | 1.054 × 10−6 | 8.654 × 10−9 |
4 | 2.877 × 10−5 | 1.336 × 10−6 | 9.988 × 10−9 | 3.359 × 10−12 |
6 | 2.915 × 10−5 | 3.111 × 10−7 | 2.819 × 10−10 | 3.874 × 10−15 |
x | σ = 1/4 | σ = 1/3 | σ = 2/5 | σ = 2/3 |
---|---|---|---|---|
4 | 5.260 × 10−2 | 3.447 × 10−4 | 6.825 × 10−2 | 6.130 × 10−4 |
6 | 2.176 × 10−4 | 1.570 × 10−5 | 2.863 × 10−2 | 2.988 × 10−6 |
8 | 6.088 × 10−6 | 2.510 × 10−6 | 5.153 × 10−4 | 3.365 × 10−9 |
10 | 3.787 × 10−6 | 3.111 × 10−7 | 4.993 × 10−5 | 6.279 × 10−11 |
12 | 1.048 × 10−7 | 1.508 × 10−8 | 1.431 × 10−7 | 2.397 × 10−12 |
Then from Theorem 2.2 we obtain the leading behavior:
as x → +∞ and ϵ → 0. The above approximation for Mσ(x) agrees with that obtained in [6] by application of the saddle-point method applied to the integral (1.2). This argument is explained in Section 5.
Plots of Mσ(x) given by (4.1) are shown in Figs. 1, 2 and 3, and plots of Mσ(−x) given by (4.2) are shown in Fig. 4. These illustrate the transition to a Dirac delta function as ϵ → 0.
5 The Kreis-Pipkin Method
This section focuses on the argument introduced as a variant of the saddle-point method by Kreis and Pipkin in [2] (revisited by Mainardi and Tomirotti in [7] for a wave problem in fractional viscoelasticity) to deal with sharply peaked functions around x ~ 1, in the limit where ϵ → 0+. The method is of interest from a numerical point of view, allowing us to deal with functions that are also physically relevant such as, in seismology, the pulse response in the nearly elastic limit.
In this way it is possible, adapting the Kreis-Pipkin method to the M-Wright function, to study its asymmetric structure when it tends towards the Dirac delta function δ(x − 1).
We start by recalling the integral definition of the auxiliary Wright function Fσ(x) (compare (1.2))
related to the function Mσ(x) by (1.5). Taking into account the procedure described in [2], we have with σ = 1 − ϵ that the exponent is stationary at the point:
The next step is to expand t−ϵ in powers of ϵln t/t0, this being more accurate than expanding the exponent in powers of t − t0, and using z = t/t0. The final result is:
where we emphasize that this procedure is valid only in the limit ϵ → 0+. The relation (1.5) tells us that the expression of Mσ(x) can be simply obtained from knowledge of Fσ(x), and vice versa. The exponential factor appearing in (5.2) has a saddle point at z = 1 and the contour can be made to coincide with the steepest descent path, which is locally perpendicular to the real z-axis at the saddle. Then finally, by means of the steepest descent method, the function Mσ(x) as σ → 1− can be expressed via a real integral.
The results are presented in Figs. 5, 6 and 7; each figure shows a comparison in linear and semi-logarithmic scale between three curves obtained using different methods. These are respectively the Kreis-Pipkin method, (4.1) of this work and the classical saddle-point method used by Mainardi and Tomirotti [6] (denoted by M-T 1995 in the figures). Note that the curves obtained via (4.1) and M-T 1995 are equivalent, and indeed can be simply shown to be analytically equivalent. The plots for 0 ≤ x ⋍ 1 in the Kreis-Pipkin method were obtained via an integral representation for Mσ(x) combined with matching to the leading asymptotic behavior. The method proposed by Kreis and Pipkin is thus seen to be a useful tool to reproduce the asymmetric structure of Mσ(x) that would be impossible with the standard saddle-point method.
6 Conclusions
We have given asymptotic expansions as x → ±∞ for the auxiliary Wright functions Fσ(x) and Mσ(x) defined in (1.3) and (1.4) when 0 < σ < 1. These expansions consist of series of an exponential and algebraic character whose relative dominance depends on the parameter σ. An algorithm for determining the coefficients in the exponential expansion is discussed and explicit representation of the first few coefficients has been given.
Numerical results are presented to confirm the accuracy of the expansions. Of particular interest is the the limit σ → 1−, where the function Mσ(x) approaches a Dirac delta function centered on x = 1. Graphical results based on the Kreis-Pipkin method are given that illustrate the leading asymptotic forms and the transition of Mσ(x) to a delta function.
Appendix A An algorithm for the computation of the coefficients cj(σ)
In this Appendix we describe an algorithm for the computation of the coefficients cj(σ) appearing in the exponential expansion of the function
for |s| → ∞ uniformly in |arg s| ≤ π − ϵ, where the parameters κ, h, ϑ, A0(σ) are defined in (2.1), with ϑ′ = 1 − ϑ.
Introduction of the scaled Gamma function
where
Then, after some routine algebra we find that the left-hand side of (A.1) can be written as
where
Substitution of (A.2) in (A.1) then yields the inverse factorial expansion in the alternative form
as |s| → ∞ in |arg s| ≤ π − ϵ.
We now expand R(s) and ϒ(s) for s → +∞ making use of the well-known expansion (see, for example, [12, p. 71])
where γk are the Stirling coefficients with γ0 = 1,
whence
where we have defined the quantities
Upon equating coefficients of s−1 in (A.3) we then obtain
The higher coefficients are obtained by continuation of this expansion process in inverse powers of s. We write the product on the left-hand side of (A.3) as an expansion in inverse powers of κs in the form
as s → +∞, where the coefficients Cj are determined with the aid of Mathematica; see [10, Appendix A] for details. The coefficients cj(σ) are then obtained by a recursive process to yield the expressions given in (2.4). This procedure is found to work well in specific cases when the various parameters have numerical values, where up to a maximum of 100 coefficients have been so calculated.
Acknowledgements
The research activity of AC and FM has been carried out in the framework of the activities of the National Group of Mathematical Physics (GNFM, INdAM). The activity of AC, PhD student at the University of Würzburg, is carried out also in the Würzburg-Dresden Cluster of Excellence - Complexity and Topology in Quantum Matter (ct.qmat).
References
[1] D. O. Cahoy, (2012) Estimation and simulation for the M-Wright function. Communications in Statistics - Theory and Methods41, No 8 (2012), 1466–1477; DOI:10.1080/03610926.2010.543299.Search in Google Scholar
[2] A. Kreis and A.C. Pipkin, : Viscoelastic pulse propagation and stable probability distributions. Quart. Appl. Math. 44 (1986), 353–360.10.1090/qam/856190Search in Google Scholar
[3] V. Lipnevich and Yu. Luchko, The Wright function: Its properties, applications, and numerical evaluation. AIP Conference Proceedings1301 (2010), Art. 614; (2010) DOI: 10.1063/1.3526663.Search in Google Scholar
[4] F. Mainardi, Fractional Calculus and Waves in Linear Viscoelasticity. Imperial College Press, London, 2010.10.1142/p614Search in Google Scholar
[5] F. Mainardi and A. Consiglio, The Wright function of the second kind in mathematical physics. In: “SI on Special Functions with Applications in Mathematical Physics”, Mathematics8, No 6 (2020), Art. 884.10.3390/math8060884Search in Google Scholar
[6] F. Mainardi and M. Tomirotti, On a special function arising in the time fractional diffusion-wave equation. In: P. Rusev, I. Dimovski and V. Kiryakova (Eds), Transform Methods and Special Functions, Sofia 1994, Science Culture Technology, Singapore (1995), 171–183. [Proc. Int. Workshop Transform Methods and Special Functions, Sofia 12–17 August 1994].Search in Google Scholar
[7] F. Mainardi and M. Tomirotti, Seismic pulse propagation with constant Q and stable probability distributions. Annali di Geofisica40 (1997), 1311–1328; E-print: http://arxiv.org/abs/1008.1341.10.4401/ag-3863Search in Google Scholar
[8] F.W.J. Olver, D.W. Lozier, R.F. Boisvert and C.W. Clark (Editors), NIST Handbook of Mathematical Functions. Cambridge University Press, Cambridge, 2010.Search in Google Scholar
[9] R.B. Paris, Exponentially small expansions of the Wright function on the Stokes lines. Lithuanian Math. J. 54 (2014), 82–105.10.1007/s10986-014-9229-9Search in Google Scholar
[10] R.B. Paris, The asymptotics of the generalised Bessel function. Math. Aeterna7 (2017), 381–406.Search in Google Scholar
[11] R.B. Paris, Asymptotics of the special functions of fractional calculus. In: A. Kochubei, Y. Luchko (Eds). Handbook of Fractional Calculus with Applications, Vol. 1, 297–325, De Gruyter, Berlin, 2019.10.1515/9783110571622-012Search in Google Scholar
[12] R.B. Paris and D. Kaminski, Asymptotics and Mellin-Barnes Integrals. Cambridge University Press, Cambridge, 2001.10.1017/CBO9780511546662Search in Google Scholar
[13] R.B. Paris and V. Vinogradov, Asymptotic and structural properties of the Wright function arising in probability theory. Lithuanian Math. J. 56 (2016), 377–409.10.1007/s10986-016-9324-1Search in Google Scholar
[14] E.M. Wright, The asymptotic expansion of the generalized Bessel function. Proc. Lond. Math. Soc. (Ser. 2)38 (1934), 286–293.10.1112/plms/s2-38.1.257Search in Google Scholar
[15] E.M. Wright, The generalized Bessel function of order greater than one. Quart. J. Math. 11 (1940), 36–48.10.1093/qmath/os-11.1.36Search in Google Scholar
© 2021 Diogenes Co., Sofia