Exact solutions of triple-order time-fractional differential equations for anomalous relaxation and diffusion I: The accelerating case

https://doi.org/10.1016/j.physa.2010.10.012Get rights and content

Abstract

In recent years the interest around the study of anomalous relaxation and diffusion processes is increased due to their importance in several natural phenomena. Moreover, a further generalization has been developed by introducing time-fractional differentiation of distributed order which ranges between 0 and 1. We refer to accelerating processes when the driving power law has a changing-in-time exponent whose modulus tends from less than 1 to 1, and to decelerating processes when such an exponent modulus decreases in time moving away from the linear behaviour. Accelerating processes are modelled by a time-fractional derivative in the Riemann–Liouville sense, while decelerating processes by a time-fractional derivative in the Caputo sense. Here the focus is on the accelerating case while the decelerating one is considered in the companion paper. After a short reminder about the derivation of the fundamental solution for a general distribution of time-derivative orders, we consider in detail the triple-order case for both accelerating relaxation and accelerating diffusion processes and the exact results are derived in terms of an infinite series of H-functions. The method adopted is new and it makes use of certain properties of the generalized Mittag-Leffler function and the H-function, moreover it provides an elegant generalization of the method introduced by Langlands (2006) [T.A.M. Langlands, Physica A 367 (2006) 136] to study the double-order case of accelerating diffusion processes.

Research highlights

► Exact solution triple-order time-fractional anomalous relaxation. ► Exact solution triple-order time-fractional anomalous diffusion. ► New method to solve multi-order time-fractional differential equations. ► Prabhakar generalized Mittag-Leffler function.

Introduction

Anomalous dynamics is frequently met in processes through complex and/or disordered media, e.g. dispersion in plasmas [1] or diffusion with obstacles [2] or binding [3], self-diffusion of surfactant molecules [4] or protein movements [5], [6], and light scattering in a cold atomic cloud [7]. A useful mathematical tool for physical investigation and description of such phenomena is fractional calculus, see for example Refs. [8], [9], [10], [11], [12], [13] for anomalous relaxation and Refs. [14], [15], [16], [17], [18], [19], [20] for anomalous diffusion. Recently, the extension of fractional differential equations to distributed-order fractional differential equations has permitted to describe also processes whose scaling law changes in time [21], [22], [23], [12], [24], [25], [26], [27], [28]. However an early idea of the time-fractional derivative of distributed order was proposed in 1969 by Caputo [29], and later re-proposed by Caputo himself [30], [31] and Bagley and Torvik [32], [33]. In particular, when the time-fractional derivative operator is opportunely chosen, it permits to model phenomena whose driving power law can be expressed by a growing-in-time exponent modulus that from less than 1 tends to 1, or an exponent modulus that decreases in time moving away from the linear behaviour, and we refer to them as accelerating and decelerating processes, respectively [34], [35]. The accelerating processes are modelled by a Riemann–Liouville (R–L) time-fractional derivative while the decelerating processes by a Caputo (C) time-fractional derivative. However, when single-order fractional differential equations are considered the two forms (R–L) and (C) are equivalent. The present paper is focused on the accelerating case while the decelerating case is considered in the companion paper [36].

A double-order time-fractional diffusion equation has been exactly solved by Langlands [37], [38]. The main object of the present article is to further include one more fractional time derivative and investigate its solution by the application of the generalized Mittag-Leffler function. Then the triple-order time-fractional differential equations considered in the present paper are: dudt=λ[PtD1α+QtD1β+TtD1γ]u(t),tR0+,ut=[PtD1α+QtD1β+TtD1γ]2x2u(x,t),tR0+,xR, with 0<α<β<γ<1, where λ is a positive constant and tDν is the Riemann–Liouville time-fractional derivative operator of order ν>0 [39], [40], which for a sufficiently well-behaved function f(t) is defined as tDνf(t)={dmdtm[1Γ(mν)0tf(τ)dτ(tτ)ν+1m],m1<ν<m,dmdtmf(t),ν=m.

The rest of the paper is organized as follows. In Section 2 the fundamental solution for a general distribution of time-derivative orders is recalled together with general considerations which motivate the name accelerating. In Section 3 the exact solutions for both triple-order time-fractional relaxation and diffusion equations are obtained using a new method based on certain properties of Mittag-Leffler and H-functions. Concluding remarks are given in Section 4.

Section snippets

Accelerating time-fractional relaxation equation

The equation of accelerating time-fractional relaxation is dudt=λ01p(ν)tD1νu(t)dν,u(0+)=1,λR+,tR0+, where p(ν)0 is the weight function of the fractional order derivative, which is taken normalized; i.e. 01p(ν)dν=1. A general theoretical analysis of time-fractional relaxation of distributed order can be found in Ref. [12]. Let the Laplace transform for a generic function w(t) be defined as: L{w(t);s}=w˜(s)0+estw(t)dt,sC. We recall that, if the limiting values of the k-integer

Solutions of triple-order accelerating relaxation and diffusion

In this section we present a new method to calculate the exact solution of the triple-order time-fractional differential equation of accelerating relaxation (1) and accelerating diffusion (2), which is based on the Prabhakar generalization of the Mittag-Leffler function, see Appendix B. In general, the self-similarity of solutions of the ordinary single-order time-fractional equations is due to the unique derivative order, so that such self-similarity is lost in distributed cases. In

Conclusion

In the present paper we have considered the triple-order time-fractional differential equations, with derivative orders less than 1, for modelling both accelerating relaxation and accelerating diffusion, by using Riemann–Liouville fractional differential operator. The corresponding analysis for decelerating relaxation and decelerating diffusion, by using Caputo fractional differential operator, is considered in the companion paper [36].

A new method is outlined. It requires certain properties of

Acknowledgements

The authors would like to thank Prof. F. Mainardi for comments and suggestions and the anonymous referees who highlighted to us formula (B.4).

References (58)

  • M. Saxton

    Biophys. J.

    (1994)
  • M. Saxton

    Biophys. J.

    (1996)
  • A. Reynolds

    Phys. Lett. A

    (2005)
  • M. Saxton

    Biophys. J.

    (2001)
  • F. Mainardi

    Chaos Solitons Fractals

    (1996)
  • L.-J. Lv et al.

    J. Comput. Appl. Math.

    (2009)
  • R. Metzler et al.

    Phys. Rep.

    (2000)
  • A. Piryatinska et al.

    Physica A

    (2005)
  • G.M. Zaslavsky

    Phys. Rep.

    (2002)
  • F. Mainardi et al.

    Appl. Math. Comput.

    (2007)
  • F. Mainardi et al.

    J. Comput. Appl. Math.

    (2007)
  • H.G. Sun et al.

    Physica A

    (2010)
  • T.A.M. Langlands

    Physica A

    (2006)
  • F. Mainardi et al.

    Appl. Math. Comput.

    (2003)
  • F. Mainardi et al.

    J. Comput. Appl. Math.

    (2005)
  • H.M. Srivastava et al.

    Appl. Math. Comput.

    (2009)
  • S. Ratynskaia et al.

    Phys. Rev. Lett.

    (2006)
  • Y. Gambin et al.

    Phys. Rev. Lett.

    (2005)
  • G. Labeyrie et al.

    Phys. Rev. Lett.

    (2003)
  • M. Caputo et al.

    Riv. Nuovo Cimento Ser. II

    (1971)
  • R. Gorenflo et al.
  • R. Hilfer
  • F. Mainardi et al.

    J. Vib. Control

    (2007)
  • T.F. Nonnenmacher et al.

    Fractals

    (1995)
  • F. Mainardi et al.

    Fract. Calc. Appl. Anal.

    (2001)
  • R. Metzler et al.

    J. Phys. A: Math. Gen.

    (2004)
  • R. Metzler et al.

    Phys. Rev. E

    (1998)
  • A.V. Chechkin et al.

    Fract. Calc. Appl. Anal.

    (2003)
  • F. Mainardi et al.
  • Cited by (23)

    • Finiteness conditions for performance indices in generalized fractional-order systems defined based on the regularized Prabhakar derivative

      2023, Communications in Nonlinear Science and Numerical Simulation
      Citation Excerpt :

      A class of the well-known fractional-order operators is Prabhakar fractional operators. In 1971, Tilak Raj Prabhakar [3] introduced the three-parameter generalized Mittag-Leffler function (T-PGMLF) which plays a significant role in fractional calculus and has attracted increasing interest in several areas, e.g. in the stochastic processes [4], probability theory [5], physics [6], systems with strong anisotropy [7] and in the description of dynamical models of spherical stellar systems [8]. Some important advantages of T-PGMLF are reported in [9–11].

    • A graphical tuning method for fractional order controllers based on iso-slope phase curves

      2020, ISA Transactions
      Citation Excerpt :

      It was first established together with integer calculus, and afterwards rediscovered in the last decades in order to get a better mathematical description of the environment. Then, it has been found useful in many fields, from Economics [1] to Physics [2], and therefore, Engineering. Modeling [3–5] and system control [6], specially robust control, as in [7] and [8], are the main applications of fractional calculus in Engineering.

    • On the time-fractional Cattaneo equation of distributed order

      2019, Physica A: Statistical Mechanics and its Applications
    • Beyond monofractional kinetics

      2017, Chaos, Solitons and Fractals
    • Generalized distributed order diffusion equations with composite time fractional derivative

      2017, Computers and Mathematics with Applications
      Citation Excerpt :

      Generalization of time fractional and distributed order time fractional diffusion equations can be done in the framework of the CTRW theory by introduction of generalized waiting time PDF [3] with a given non-negative integrable function, which appears as a memory kernel from the left hand side in the diffusion equation. Such distributed order diffusion equations have been shown to represent useful tool for modeling decelerating anomalous diffusion, ultraslow diffusive processes and strong anomaly [9–19]. These equations have been recently shown to possess multiscaling properties [17,18].

    View all citing articles on Scopus
    View full text