(Download PDF) Special Functions of Fractional Calculus Applications To Diffusion and Random Search Processes Trifce Sandev Full Chapter PDF

Special Functions of Fractional
Calculus: Applications to Diffusion and

Random Search Processes Trifce
Sandev
Visit to download the full and correct content document:
https://ebookmass.com/product/special-functions-of-fractional-calculus-applications-to
-diffusion-and-random-search-processes-trifce-sandev/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Fractional Calculus and Fractional Processes with

Applications to Financial Economics: Theory and
Application 1st Edition Fabozzi
https://ebookmass.com/product/fractional-calculus-and-fractional-
processes-with-applications-to-financial-economics-theory-and-
application-1st-edition-fabozzi/
Schaum's Outline of Probability, Random Variables, and

Random Processes, Fourth Edition Hwei P. Hsu
https://ebookmass.com/product/schaums-outline-of-probability-
random-variables-and-random-processes-fourth-edition-hwei-p-hsu/
Micro-nanoelectronics Devices : Modeling of Diffusion

and Operation Processes Gontrand
https://ebookmass.com/product/micro-nanoelectronics-devices-
modeling-of-diffusion-and-operation-processes-gontrand/
Calculus: Early Transcendental Functions 4th Edition,

(Ebook PDF)
https://ebookmass.com/product/calculus-early-transcendental-
functions-4th-edition-ebook-pdf/
Calculus Early transcendental functions 7e 2019 Rob
Larson
https://ebookmass.com/product/calculus-early-transcendental-
functions-7e-2019-rob-larson/
Spatiotemporal Random Fields, Second Edition: Theory

and Applications George Christakos
https://ebookmass.com/product/spatiotemporal-random-fields-
second-edition-theory-and-applications-george-christakos/
Stochastic processes and random matrices : lecture

notes of the Les Houches Summer School First Edition.
Edition Altland
https://ebookmass.com/product/stochastic-processes-and-random-
matrices-lecture-notes-of-the-les-houches-summer-school-first-
edition-edition-altland/
Handbook of solid state diffusion volume 1 Diffusion

Fundamentals and Techniques Divinski
https://ebookmass.com/product/handbook-of-solid-state-diffusion-
volume-1-diffusion-fundamentals-and-techniques-divinski/
Functions Modeling Change: A Preparation for Calculus,

5th Edition – Ebook PDF Version
https://ebookmass.com/product/functions-modeling-change-a-
preparation-for-calculus-5th-edition-ebook-pdf-version/
SPECIAL FUNCTIONS OF
FRACTIONAL CALCULUS
Applications to Diffusion
and Random Search Processes
12743_9789811252945_TP.indd 1 22/8/22 3:38 PM

B1948 Governing Asia
This page intentionally left blank
B1948_1-Aoki.indd 6 9/22/2014 4:24:57 PM

SPECIAL FUNCTIONS OF
FRACTIONAL CALCULUS
Applications to Diffusion
and Random Search Processes
Trifce Sandev
Macedonian Academy of Sciences and Arts, Macedonia
Alexander Iomin
Technion-Israel Institute of Technology, Israel
NEW JERSEY • LONDON • SINGAPORE • BEIJING • SHANGHAI • HONG KONG • TAIPEI • CHENNAI • TOKYO
12743_9789811252945_TP.indd 2 22/8/22 3:38 PM

Published by
World Scientific Publishing Co. Pte. Ltd.
5 Toh Tuck Link, Singapore 596224
USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601
UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE
Library of Congress Control Number: 2022945045
British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library.
SPECIAL FUNCTIONS OF FRACTIONAL CALCULUS

Applications to Diffusion and Random Search Processes
Copyright © 2023 by World Scientific Publishing Co. Pte. Ltd.
All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means,
electronic or mechanical, including photocopying, recording or any information storage and retrieval
system now known or to be invented, without written permission from the publisher.
For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance
Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy
is not required from the publisher.
ISBN 978-981-125-294-5 (hardcover)

ISBN 978-981-125-295-2 (ebook for institutions)
ISBN 978-981-125-296-9 (ebook for individuals)
For any available supplementary material, please visit

https://www.worldscientific.com/worldscibooks/10.1142/12743#t=suppl
Desk Editor: Nur Syarfeena Binte Mohd Fauzi
Printed in Singapore
Syarfeena - 12743 - Special Functions of Fractional Calculus.indd 1 25/8/2022 12:18:25 pm

August 12, 2022 16:49 ws-book9x6 Special Functions of Fractional Calculus 12743-main page v
To my wife Irina.
TS
To the memory of my father Miron-Meer Iomin,

son of Herchen and Bluma, may his memory be blessed.
AI
v
B1948_1-Aoki.indd 6 9/22/2014 4:24:57 PM

August 29, 2022 10:40 ws-book9x6 Special Functions of Fractional Calculus 12743-main page vii
Preface
Мы все учились понемногу,

Чему-нибудь и как-нибудь1
А. С. Пушкин
“Евгений Онегин”
Mathematics is the art of giving things misleading names.

The beautiful – and at first look mysterious – name the
fractional calculus is just one of the those misnomers
which are the essence of mathematics.2
I. Podlubny
Investigation of stochastic processes in complex media has been at-

tracted much attention for years. Theoretical modeling of diffusion in
heterogeneous and disordered media is a considerable part of these stud-
ies. Heterogeneous and disordered materials include various materials with
defects, multi-scale amorphous composites, fractal and sparse structures,
weighted graphs and networks, and so on. Diffusion in such complicated
media with geometric constraint and random forces is often anomalous.
Further developing of the theory for understanding and description of these
random processes in a variety of realizations in physics, biology, social life
and finance is a part of modern studies, what we called “complex systems”.
These complex phenomena entailing complex descriptions by fractional ki-
netics and fractional calculus.
The new mathematical approach sheds light on many questions are still ex-
ist for studying and opens new ones. Such an example is a random search
1 We all learned a little, Anything and somehow (A. S. Pushkin, “Eugene Onegin”).
2 Ref. [13], Chapter 1.
vii
August 12, 2022 16:49 ws-book9x6 Special Functions of Fractional Calculus 12743-main page viii
viii Special Functions of Fractional Calculus
process whose systematic research stems from projects involving hunting for
submarines during World War II, while the modern study of first-passage,
or hitting times covers a large area of search problems from animal food for-
aging to molecular reactions and gene regulation. In many cases, an intro-
duction of a stochastic resetting in complex systems, has significant effects
on the first-passage properties. Moreover, the random search processes in
complex networks are important for various reasons, such as understand-
ing animal food search strategies and improving web search engines, or for
prolonging or speeding up the survival time in first-encounter tasks.
Many of the aforementioned processes can be described by various form
of random walk models, including fractional Fokker-Planck equations and
generalised Langevin equations, which in their turn describe a variety of
completely different problems shared common features. In particular, a
class of diffusion in heterogeneous environment is closely related to turbu-
lent diffusion represented by an inhomogeneous advection-diffusion equa-
tion, and it also relates to generalized geometric Brownian motion, used to
model stock prices.
The suggested book addresses technical issues surrounding the concept of
special functions of anomalous transport, or fractional kinetics, described
in the framework of fractional calculus. Much attention is paid to technical
details in using the Fox H-functions and the Mittag-Leffler functions as
well as functions related to them in description of fractional transport in
random composite media. The latter defines an extremely large class of
problems in life science and physics, medical physics and technology. We
attempt to present the material in a self-contained way, such that the book
can be a convenient reading not only for experts in the field but also for
newcomers and undergraduate students interesting in the field of fractional
calculus and fractional kinetics.
Skopje, Macedonia Trifce Sandev,

Haifa, Israel Alexander Iomin
April, 2022
August 12, 2022 16:49 ws-book9x6 Special Functions of Fractional Calculus 12743-main page ix
Acknowledgments
It is our great pleasure to thank our colleagues for collaboration, nu-

merous helpful and stimulating discussions. Their constant interest, sup-
port and encourage cannot be overestimated. Our thanks to Emad Awad
(Alexandria), Lasko Basnarkov (Skopje), Aleksei Chechkin (Potsdam &
Kharkov), Weihua Deng (Lanzhou), Viktor Domazetoski (Skopje), Sergei
Fedotov (Manchester), Katarzyna Górska (Kraków), Andrzej Horzela
(Kraków), Holger Kantz (Dresden), Ljupco Kocarev (Skopje), Ervin
Kaminski Lenzi (Ponta Grossa), Vicenç Méndez (Barcelona), Ralf Met-
zler (Potsdam), Alexander Milovanov (Rome), Arnab Pal (Chennai), Irina
Petreska (Skopje), Haroldo Valentin Ribeiro (Maringa), R. K. Singh (Bar-
Ilan), Igor M. Sokolov (Berlin), Viktor Stojkoski (Skopje), Živorad To-
movski (Ostrava).
TS sincerely thanks Academician Ljupco Kocarev, President of the
Macedonian Academy of Sciences and Arts, for the given trust, great collab-
oration and continuous support in my research all these years, and for pro-
viding a very professional and positive working atmosphere at the Academy.
TS expresses his special thanks to Prof. Dr. Ralf Metzler for the un-
forgettable moments in doing research in his group at the University of
Potsdam, where a part of this book has been written, and for the great
collaboration and support.
TS acknowledges DFG (Deutsche Forschungsgemeinschaft) research
grant (DFG, Grant number ME 1535/12-1) for the project “Diffusion and
random search in heterogeneous media: theory and applications” between
Macedonian Academy of Sciences and Arts and University of Potsdam.
TS acknowledges financial support by the Alexander von Humboldt
(AvH) Foundation of the project “Statistical Mechanics of Diffusion Pro-
cesses in Disordered Media” (2020–2022), within the AvH fellowship
ix
August 12, 2022 16:49 ws-book9x6 Special Functions of Fractional Calculus 12743-main page x
x Special Functions of Fractional Calculus
programme for experienced researchers at the Institute of Physics & As-

tronomy, University of Potsdam.
TS also acknowledges support from the bilateral Macedonian-Chinese
research project 20-6333, between the Macedonian Academy of Sciences
and Arts and Lanzhou University, funded under the intergovernmental
Macedonian-Chinese agreement.
AI sincerely thanks Distinguished Professor Mordechai (Moti) Segev for
his continuous support.
AI also acknowledges financial support by the Israel KAMEA Program.
The authors thank the Editorial Consultant S. C. Lim and the Edi-
tor Mrs. Nur Syarfeena Binte Mohd Fauzi for the collaboration and their
management of the publication process of this book.
Last, but not least, we are thankful to our families for all the patience
and support in our work.
August 12, 2022 16:49 ws-book9x6 Special Functions of Fractional Calculus 12743-main page xi
Contents
Preface vii
Acknowledgments ix
Acronyms and Symbols xv
1. Mathematical background 1
1.1 Integral transforms . . . . . . . . . . . . . . . . . . . . . . 1
1.1.1 Fourier transform . . . . . . . . . . . . . . . . . . 2
1.1.2 Laplace transform . . . . . . . . . . . . . . . . . . 5
1.1.3 Mellin transform . . . . . . . . . . . . . . . . . . . 10
1.2 Asymptotic expansions . . . . . . . . . . . . . . . . . . . . 13
1.2.1 Tauberian theorems . . . . . . . . . . . . . . . . . 13
1.2.2 Generating function formalism . . . . . . . . . . . 15
1.2.3 Laplace method . . . . . . . . . . . . . . . . . . . 16
1.3 Special functions . . . . . . . . . . . . . . . . . . . . . . . 18
1.3.1 Gamma function . . . . . . . . . . . . . . . . . . . 18
1.3.2 Contour integral representation of Γ(p) and 1/Γ(p) 21
1.3.3 Mittag-Leffler functions . . . . . . . . . . . . . . . 22
1.4 Fractional calculus . . . . . . . . . . . . . . . . . . . . . . 34
1.4.1 Fractional integrals . . . . . . . . . . . . . . . . . 36
1.4.2 Fractional derivatives . . . . . . . . . . . . . . . . 36
1.4.3 Generalized operators . . . . . . . . . . . . . . . . 38
2. Fox H-function and related functions 45

2.1 Introduction: F and G higher transcendental functions . . 45
2.1.1 Meijer G-functions . . . . . . . . . . . . . . . . . . 46
xi
August 12, 2022 16:49 ws-book9x6 Special Functions of Fractional Calculus 12743-main page xii
xii Special Functions of Fractional Calculus
2.2 Fox H-function . . . . . . . . . . . . . . . . . . . . . . . . 49

2.2.1 Properties of the Fox H-function . . . . . . . . . . 50
2.2.2 Integral transforms of the Fox H-function . . . . . 54
2.2.3 Series and asymptotic expansions . . . . . . . . . 55
2.2.4 Relation to other functions . . . . . . . . . . . . . 56
2.2.5 Stable distributions . . . . . . . . . . . . . . . . . 61
3. Elements of random walk theory 65

3.1 Stochastic variables and characteristics . . . . . . . . . . . 65
3.1.1 Averaged characteristics . . . . . . . . . . . . . . . 66
3.2 Markov processes . . . . . . . . . . . . . . . . . . . . . . . 67
3.2.1 Smoluchowski–Chapman–Kolmogorov equation . . 68
3.2.2 Markov chains . . . . . . . . . . . . . . . . . . . . 69
3.2.3 Steady state, spectral decomposition and KS entropy 71
3.2.4 Random Markov chain . . . . . . . . . . . . . . . 73
3.3 The first passage . . . . . . . . . . . . . . . . . . . . . . . 75
3.4 Continuous time random walk . . . . . . . . . . . . . . . . 78
3.4.1 Renewal theory . . . . . . . . . . . . . . . . . . . 78
3.4.2 Random steps . . . . . . . . . . . . . . . . . . . . 80
3.4.3 Montroll-Weiss equation . . . . . . . . . . . . . . . 80
3.5 CTRW and fractional diffusion equations . . . . . . . . . . 82
3.5.1 Diffusion limit . . . . . . . . . . . . . . . . . . . . 82
3.5.2 A general case of FFPE . . . . . . . . . . . . . . . 86
3.6 Stochastic equations . . . . . . . . . . . . . . . . . . . . . 89
3.6.1 Fokker-Planck equation . . . . . . . . . . . . . . . 90
3.7 Langevin equation: Parametrization and FFPE . . . . . . 91
3.7.1 The PDF in physical time . . . . . . . . . . . . . . 95
4. CTRW on combs 99
4.1 CTRW and subordination . . . . . . . . . . . . . . . . . . 99
4.1.1 From standard to generalized Fokker-Planck
equation . . . . . . . . . . . . . . . . . . . . . . . 100
4.2 Fractional Fokker-Planck equation . . . . . . . . . . . . . 102
4.3 Diffusion on comb structures: Derivation of the FFPE . . 102
4.3.1 Fokker-Planck equation for comb structure . . . . 103
4.3.2 Diffusion on 3D comb . . . . . . . . . . . . . . . . 107
4.3.3 From drift-diffusion to FFPE . . . . . . . . . . . . 112
August 12, 2022 16:49 ws-book9x6 Special Functions of Fractional Calculus 12743-main page xiii
Contents xiii
5. Heterogeneous diffusion processes 119

5.1 Heterogeneous diffusion . . . . . . . . . . . . . . . . . . . 119
5.1.1 Stratonovich interpretation . . . . . . . . . . . . . 122
5.1.2 Isothermal Klimontovich-Hänggi interpretation . . 122
5.1.3 Itô interpretation . . . . . . . . . . . . . . . . . . 123
5.2 Turbulent diffusion and geometric Brownian motion . . . 124
5.2.1 Geometric Brownian motion and turbulent
diffusion . . . . . . . . . . . . . . . . . . . . . . . 124
5.2.2 Turbulent diffusion due to comb geometry . . . . 128
5.2.3 Solution of turbulent diffusion equation on comb . 133
5.2.4 Generalized geometric Brownian motion . . . . . . 134
6. Diffusion processes with stochastic resetting 141

6.1 Introduction to diffusion with Poissonian resetting . . . . 141
6.1.1 Renewal equation . . . . . . . . . . . . . . . . . . 142
6.2 Fokker-Planck equation for diffusion with resetting . . . . 143
6.2.1 Subordination approach . . . . . . . . . . . . . . . 144
6.2.2 Numerical simulations: Langevin equation
approach . . . . . . . . . . . . . . . . . . . . . . . 146
6.2.3 Transition to the steady state . . . . . . . . . . . 146
6.3 Diffusion-advection equation with resetting . . . . . . . . 148
6.3.1 Transition to the steady state . . . . . . . . . . . 151
6.4 Diffusion in comb with confining branches with resetting . 152
6.5 Diffusion in 3D comb with resetting . . . . . . . . . . . . . 158
7. Random search 171

7.1 One dimensional search . . . . . . . . . . . . . . . . . . . 171
7.1.1 One dimensional Brownian search with drift . . . 173
7.1.2 One dimensional turbulent diffusion search . . . . 173
7.2 Random search on comb-like structures . . . . . . . . . . . 175
7.2.1 Brownian search with drift on comb . . . . . . . . 175
7.2.2 Random search on three dimensional comb . . . . 179
7.2.3 Turbulent diffusion search on comb . . . . . . . . 181
8. Diffusion on fractal tartan 185

8.1 Grid comb model . . . . . . . . . . . . . . . . . . . . . . . 186
8.1.1 Finite number of backbones . . . . . . . . . . . . . 186
8.1.2 Fractal grid: infinite number of backbones . . . . 190
August 12, 2022 16:49 ws-book9x6 Special Functions of Fractional Calculus 12743-main page xiv
xiv Special Functions of Fractional Calculus
8.2 Fractal mesh . . . . . . . . . . . . . . . . . . . . . . . . . 194

8.2.1 From Weierstrass function to the power law . . . . 197
8.3 Fractal structure of fingers . . . . . . . . . . . . . . . . . . 202
8.4 Superdiffusion due to compensation kernel . . . . . . . . . 205
9. Finite-velocity diffusion 211

9.1 Cattaneo equation . . . . . . . . . . . . . . . . . . . . . . 212
9.2 Cattaneo equation on a comb structure . . . . . . . . . . . 215
9.2.1 Infinite domain solution . . . . . . . . . . . . . . . 215
9.2.2 Fractional Cattaneo equation . . . . . . . . . . . . 217
9.2.3 Finite domain solution . . . . . . . . . . . . . . . 219
9.3 Persistent random walk . . . . . . . . . . . . . . . . . . . 222
10. Appendices 229
Appendix A Functional calculus 231

A.1 Functional derivative . . . . . . . . . . . . . . . . . . . . . 231
A.2 Functional integral . . . . . . . . . . . . . . . . . . . . . . 234
Appendix B Stochastic differential equations 237

B.1 Itô stochastic calculus . . . . . . . . . . . . . . . . . . . . 237
B.2 Fokker-Planck equation . . . . . . . . . . . . . . . . . . . 239
B.2.1 Fokker-Planck equation – Itô interpretation . . . . 240
B.2.2 Fokker-Planck equation – Stratonovich
interpretation . . . . . . . . . . . . . . . . . . . . 240
Appendix C Large deviation principle 243

C.1 Large deviation function . . . . . . . . . . . . . . . . . . . 243
C.2 Legendre transform . . . . . . . . . . . . . . . . . . . . . . 244
Appendix D Fractals and fractal dimension 247

D.1 Fractal diffusion and dynamical dimensions . . . . . . . . 248
Appendix E Implementation of Wolfram Mathematica 251

E.1 Numerical inversion of Laplace transform . . . . . . . . . 253
Index 267
August 12, 2022 16:49 ws-book9x6 Special Functions of Fractional Calculus 12743-main page xv
Acronyms and Symbols
CTRW continuous time random walk

FATD first arrival time distribution
FPE Fokker-Planck equation
FFPE fractional Fokker-Planck equation
GBM geometric Brownian motion
M-L Mittag-Leffler
MSD mean squared displacement
PDF probability density function
R-L Riemann-Liouville
lhs left-hand side
rhs right-hand side
T integral transform
−1
T inverse integral transform
F Fourier transform
F −1 inverse Fourier transform
L Laplace transform
−1
L inverse Laplace transform
M Mellin transform
M−1 inverse Mellin transform
D diffusion coefficient
Γ(z) gamma function
B(a, b) beta function
δ(z) Dirac delta function
Θ(z) Heaviside step function
Γ(ν, z) upper incomplete gamma function
γ(ν, z) lower incomplete gamma function
(γ)k Pochhammer symbol
xv
August 12, 2022 16:49 ws-book9x6 Special Functions of Fractional Calculus 12743-main page xvi
xvi Special Functions of Fractional Calculus
erf(z) error function

erfc(z) complementary error function
Ēn (z) generalized exponential integral function
Eα (z) one parameter Mittag-Leffler function
Eα,β (z) two parameter Mittag-Leffler function
γ
Eα,β (z) three parameter Mittag-Leffler function
Eα (t; λ) associated one parameter Mittag-Leffler function
Eα,β (t; λ) associated two parameter Mittag-Leffler function
γ
Eα,β (t; λ) associated three parameter Mittag-Leffler function
m,n
Hp,q (z) Fox H-function
µ
Ia+ Riemann-Liouville fractional integral
µ µ
RL Da+ , Da+ Riemann-Liouville fractional derivative
µ
TRL Da+ tempered Riemann-Liouville fractional derivative
µ
C Da+ Caputo fractional derivative
µ
TC Da+ tempered Caputo fractional derivative
µ
W D−∞ Weyl fractional derivative
µ
RF Dθ asymmetric Riesz-Feller derivative
µ
α
RF D0 symmetric Riesz-Feller derivative
∂ α
∂|x|α , ∂|x| Riesz fractional derivative
γ
Eρ,µ,ω,a+ Prabhakar integral
γ γ,µ
Dρ,µ,ω,a+ , RL Dρ,ω,a+ Prabhakar derivative
γ γ,µ
C Dρ,µ,ω,a+ , C Dρ,ω,a+ regularized Prabhakar derivative
γ,µ
TRL Dρ,ω,a+ tempered Prabhakar derivative
γ,µ
TC Dρ,ω,a+ tempered regularized Prabhakar derivative
RL Gη,t generalized derivative operator in the Riemann-
Liouville form
C Gγ,t generalized derivative operator in the Caputo form
ω;γ,κ
Ea+;α,β generalized integral operator
Lα (z) one-sided Lévy stable probability density
ϕ(α, β; z) Wright function
p Ψq (z) Fox-Wright function
h·i ensemble average
dze ceiling function
bzc floor function
Υ(x, t) jump probability density function
ψ(t) waiting time probability density function
w(x) jump length probability density function
℘fa (t) first arrival time distribution
August 12, 2022 16:49 ws-book9x6 Special Functions of Fractional Calculus 12743-main page xvii
Acronyms and Symbols xvii
P
search reliability
E
search efficiency
B(t)
standard Brownian motion
h(u, t)
subordination function
<(z)real part of a complex number z
=(z)imaginary part of a complex number z
√
ı = −1 imaginary unit
Q.E.D. – quod erat demonstrandum – thus it has been
demonstrated
R = R real values
R+ = R+ non-negative real values
C = C complex values
Z = Z integer numbers
B1948_1-Aoki.indd 6 9/22/2014 4:24:57 PM

August 12, 2022 16:49 ws-book9x6 Special Functions of Fractional Calculus 12743-main page 1
Chapter 1
Mathematical background
1.1 Integral transforms
Many linear differential and integral equations can be easily solved by em-
ploying integral transforms. This procedure maps a problem in task into
the form of algebraic equations, which can be easily solved. A generic form
of the integral transform reads
Z b
T [f (z)](ξ) = K(ξ, z) f (z) dz = F (ξ), (1.1)
a
where K(ξ, z) is the kernel function, and the limits of the integration, (a, b),
are determined by the problem in task. The kernel function establishes a
relation between function f (z), defined in z space, and its image F (ξ) in
its ξ space, and together with limits of the integration these determine the
type of the transform. The inverse integral transformation reads
Z v
T −1 [F (ξ)](z) = K −1 (z, ξ) F (ξ) dξ, (1.2)
u
−1
where K (ξ, z) is the inverse kernel function, and u and v are the limits
of integration. An appropriate form of the integral transform is determined
by equations in task and corresponding boundary and initial conditions.
The basic property of the integral transforms is their linearity,
Z b
T [c1 f1 (z) + c2 f2 (z)](ξ) = K(ξ, z) [c1 f1 (z) + c2 f2 (z)] dz
a
Z b Z b
= c1 K(ξ, z) f1 (z) dz + c2 K(ξ, z) c2 f2 (z) dz
a a
= c1 F1 (ξ) + c2 F2 (ξ), (1.3)
where c1,2 are given constants.
1
2 Special Functions of Fractional Calculus
In this book we will employ three forms of the integral transform:

Fourier, Laplace and Mellin transforms, which are determined by the cor-
responding kernels of the transformation and the boundaries, as shown in
Table 1.1.
Table 1.1 Table of integral transforms.
Transform K(ξ, z) a b K −1 (z, ξ) u v
1 ıtξ
F e−ıξz −∞ ∞ 2π
e −∞ ∞
1 ξz
L e−ξ z 0 ∞ 2πı
e c − ı∞ c + ı∞
1
M z ξ−1 0 ∞ 2πı
z −ξ c − ı∞ c + ı∞
1.1.1 Fourier transform
As shown in Table 1.1, the kernel of the Fourier transform is exponential

K(k, x) = e−ıkx , and taking the limits at infinity, a = −∞ and b = ∞, we
have
Z ∞
F[f (x)](k) = F (k) = f (z) e−ıkx dx, (1.4)
−∞
while the inverse Fourier transform reads

Z ∞
−1 1
F [F (k)](x) = f (x) = F (k) eıkx dk. (1.5)
2π −∞
For applications, we also use “tilde” sign for the Fourier image, f˜(k) ≡ F (k).
In addition to the basic property (1.3), the Fourier transform has the
following properties:
(1) Translation:
F[f (x − x0 )] = e−ıkx0 F (k) (1.6)

Mathematical background 3
This can be easily shown from the definition of the Fourier transform,
Z ∞ Z ∞
−ıkx
F[f (x − x0 )] = f (x − x0 )e dx = f (z)e−ık(z+x0 ) dz
−∞ (z=x−x0 ) −∞
Z ∞
−ıkx0
=e f (z)e−ıkz dz = e−ıkx0 F[f (x)] = e−ıkx0 F (k).
−∞
(2) Modulation:
F eıxk0 f (x) = F (k − k0 )

(1.7)
Namely,
Z ∞
F eıxk0 f (x) = e−ı(k−k0 )x f (x) = F (k − k0 ).

−∞
(3) Scaling (a 6= 0):

1
F[f (ax)] = F (k/a) (1.8)
|a|
It directly follows from the definition of the Fourier transform,
Z ∞
F[f (ax)] = e−ıkx f (ax) dx
−∞
Z ∞
1 1
= e−ı(k/a)z f (z) dz = F (k/a).
(z=ax) |a| −∞ |a|
(4) Convolution theorem:
Z ∞
F[f ? g] = F f (ζ)g(x − ζ) dζ = F (k)G(k) (1.9)
−∞
This property results from the change of the order of integration,

Z ∞ Z ∞ Z ∞
F f (ζ)g(x − ζ) dζ = e−ıkx f (ζ)g(x − ζ) dζ dx
−∞ −∞ −∞
Z ∞ Z ∞
= f (ζ) g(x − ζ)e−ıkζ dx dζ
−∞ −∞
Z ∞ Z ∞
−ıkζ
= e f (ζ) g(x − ζ)e−ık(x−ζ) dx dζ
−∞ −∞
Z ∞ Z ∞
−ıkζ
= e f (ζ) dζ e−ıkz g(z) dz
(x−ζ=z) −∞ −∞
= F[f (x)] · F[g(x)] = F (k) · G(k).
R∞ R∞
Here we note that −∞ f (ζ) g(x − ζ) dζ = −∞ g(ζ) f (x − ζ) dζ, i.e.,
f ? g = g ? f.
(5) Transform of derivative of f (x):

n
d
F f (x) = (ık)n F (k) (1.10)
dxn
This can be shown by mathematical induction knowing that the Fourier
transform of the first derivative is
Z ∞
d df (x)
F f (x) = e−ıkx dx
dx −∞ dx
−ıkx ∞ :0 Z ∞
e−ıkx f (x) dx

= f (x) e
−∞
+ (ık)
−∞
= (ık) F (k), (1.11)
where we use integration by parts, and of the second derivative is
2 Z ∞
d d2 f (x)
F 2
f (x) = e−ıkx dx
dx −∞ dx2
df (x) −ıkx
∞ :0 Z ∞
df (x)
= e + (ık) e−ıkx dx
dx −∞ dx
−∞
2 2
= (ık) F (k) = −k F (k). (1.12)
From the definition (1.4) follows the Fourier transform of the Dirac δ-
function
Z ∞
F [δ(x)] = δ(x) e−ıkx dx = 1, (1.13)
−∞
and exponential function
h i Z ∞ 2a
F e−a|x| = e−(ıkx+a|x|) dx = , <(a) > 0, (1.14)
−∞ a2 + k 2
which means that the Fourier transform of a two-sided decaying exponential
function is a Cauchy (or Lorentzian) function.
Remark 1.1. From the definition of the Fourier transform (1.4), it follows
directly that
Z ∞ Z ∞
f (z) dx = lim f (z) e−ıkx dx = F (0) (1.15)
−∞ k→0 −∞
The Fourier transform and the inverse Fourier transform are im-
plemented in Wolfram Language as FourierTransform[f, x, k] and
InverseFourierTransform[F, k, x], respectively. Here we note that the co-
efficients in front of the integrals in Wolfram Mathematica are symmetrical
and taken to be √12π in contrast to the integrals (1.4) and (1.5), where
1
the coefficients are taken to be 1 and 2π , respectively. In both cases the
product of the coefficients in front of the integrals in the definition of the
1
Fourier and the inverse Fourier transforms should be equal to 2π .
1.1.2 Laplace transform

The Laplace transform is described by the kernel function K(s, t) = e−st ,
and the limits a = 0 and b = ∞, see Table Z
1.1, i.e.,
∞
L[f (t)](s) = F (s) ≡ fˆ(s) = f (t) e−st dt. (1.16)
0
When dealing with the Laplace transform, we suppose that the function
f (t) equals zero for t < 0, i.e., the function is always multiplied by the
Heaviside function Θ(t), which is equal to zero for t < 0. The inverse
Laplace transform reads
Z c+ı∞
1
L−1 [F (s)](t) = f (t) = F (s) est ds. (1.17)
2πı c−ı∞
Besides the linearity basic property, the Laplace transform has also the
following properties:
(1) Time shifting:

L[f (t − t0 )] = e−st0 F (s) (1.18)
Here, we again use that the function f (t − t0 ) = 0 for t < t0 , i.e., the
function f (t − t0 ) is multiplied by the Heaviside function Θ(t − t0 ). It
can be presented Z as follows
∞ Z ∞
L[f (t − t0 )] = f (t − t0 )e−st dt = f (t − t0 )e−st dt
0 t0
Z ∞
−st0
= e f (z)e−sz dz = e−st0 L[f (t)] = e−st0 F (s).
z=t−t0 0
A different result is obtained if one takes the Laplace transform of
f (t + t0 ) since Z
∞ Z ∞
L[f (t + t0 )] = f (t + t0 )e−st dt = est0 f (t + t0 )e−s(t+t0 ) dt
0 0
Z ∞
= e st0
f (z)e−sz dz
z=t+t0 t0
Z ∞ Z t0
= est0 f (z)e−sz dz −
f (z)e−sz dz
0 0
Z t0
= est0 L [f (t)] − f (z)e−sz dz .
0
Therefore, the following property holds true:

Z t0
st0 −sz
L[f (t + t0 )] = e L [f (t)] − f (z)e dz (1.19)
0
(2) Frequency shifting:

L es0 t f (t) = F (s − s0 )

(1.20)
This property immediately follows from the integration
Z ∞
L es0 t f (t) = e−(s−s0 )t f (t) dt = F (s − s0 ).

0
(3) Time scaling (a > 0):
1
L[f (at)] =
F (s/a) (1.21)
a
From the definition of the Laplace transform one finds
Z ∞
L[f (at)] = e−st f (at) dt
0
1 ∞ −(s/a)z
Z
1
= e f (z) dz = F (s/a).
(z=ax) a 0 a
(4) Transform of convolution:
Z t
L[f ? g] = L f (τ )g(t − τ ) dτ = F (s)G(s) (1.22)
0
This property follows from Dirichlet’s formula

Z b Z x Z b Z b
dx f (x, y) dy = dy f (x, y) dx, (1.23)
a a a y
which yields
Z t Z ∞ Z t
−st
L f (τ )g(t − τ ) dτ = e f (τ )g(t − τ ) dτ dt
0 0
Z ∞ Z0 ∞
= f (τ ) g(t − τ )e−st dt dτ
0 τ
Z ∞ Z ∞
−sτ
= e f (τ ) g(t − τ )e−s(t−τ ) dt dτ
0 τ
Z ∞ Z ∞
−sτ
= e f (τ ) dτ e−sz g(z) dz
(t−τ =z) 0 0
= L[f (t)] · L[g(t)] = F (s) · G(s).
Rt Rt
We also note that 0 f (τ ) g(t − τ ) dτ = 0 g(τ ) f (t − τ ) dτ , i.e., f ? g =
g ? f.
(5) Transform of derivative of f (t):

n n
d n
X
L f (t) = s F (s) − sn−k f (k−1) (0) (1.24)
dtn
k=1
From the Laplace transform of the first derivative

Z ∞
d df (t)
L f (t) = e−st dt
dt 0 dt
Z ∞
−st ∞
= f (t) e 0
+s e−st f (t) dt
0
= sF (s) − f (0), (1.25)
and the second derivative
2 Z ∞
d d2 f (t)
L 2
f (t) = e−st dt
dt 0 dt2
∞ Z ∞
df (t) −st df (t)
= e +s e−st dt
dt 0 0 dt
= s2 F (s) − sf (0) − f 0 (0), (1.26)
by mathematical induction one obtains Eq. (1.24).
(6) Transformation of integral:
Z t
1
L f (τ ) dτ = F (s) (1.27)
0 s
This property follows from Dirichlet’s formula, Eq. (1.23), which yields
Z t Z ∞ Z t
−st
L f (τ ) dτ = e f (τ ) dτ dt
0 0 0
Z ∞ Z ∞
−st
= f (τ ) e dt dτ
0 τ
Z ∞
1 1
= e−sτ f (τ ) dτ = F (s). (1.28)
s 0 s
(7) Integration in frequency domain:
Z ∞
f (t)
F (ζ) dζ = L (1.29)
s t
This property can also be shown by exchanging the order of integration,
Z ∞ Z ∞ Z ∞
f (t)
F (ζ) dζ = e−ζt f (t) dt dζ = L . (1.30)
s s 0 t
(8) Transformation of periodic function f (t + T ) = f (t), T > 0:

Z T
1
L [f (t)] = e−st f (t) dt (1.31)
1 − e−sT 0
From the straightforward calculations, we have
Z ∞ Z T Z ∞
L [f (t)] = e−st f (t) dt = e−st f (t) dt + e−st f (t) dt
0 0 T
Z T Z ∞
= e−st f (t) dt + e−s(τ +T ) f (τ + T ) dτ
0 0
Z T Z ∞
= e−st f (t) dt + e−sT e−sτ f (τ ) dτ
0 0
Z T
= e−st f (t) dt + e−sT L [f (t)] , (1.32)
0
which yields Eq. (1.31).
One can easily find the following formulas for the Laplace transform of
a Dirac δ-function, a power function and an exponential function, respec-
tively,
Z ∞
L [δ(t)] = δ(t) e−st dt = 1, (1.33)
0
Z ∞
L [tν ] = tν e−st dt = Γ(ν + 1) s−ν−1 , <(ν) > −1, (1.34)
0
Z ∞
1
L e−at = e−(s+a)t dt =

, <(s + a) > 0, (1.35)
0 s+a
where Γ(z) is a gamma function (see Sec. 1.3.1).
Remark 1.2 (Limit results). There are useful limit relations between
the function f (t) and its Laplace transform pair F (s). For example, from
the result related to the Laplace transform of the first derivative (1.25) one
finds
Z ∞
d df (t)
sF (s) = f (0) + L f (t) = f (0) + e−st dt. (1.36)
dt 0 dt
Assuming that f (t) is an analytic function (infinitely differentiable) and
integrating by parts (u = dfdt(t) , dv = e−st dt), one obtains
1 1
sF (s) = f (0) + f 0 (0) + 2 f 00 (0) + . . . . (1.37)
s s
Then using the limit s → ∞ one obtains the following useful relation
lim sF (s) = f (0+) = lim f (t) (1.38)
s→∞ t→0
which is known as the initial value theorem. Continuing the procedure, one
obtains
lim s2 F (s) − sf (0) = f 0 (0).

(1.39)
s→∞
In contrary, in the limit s → 0, Eq. (1.36) yields
Z ∞
df (t)
lim sF (s) = f (0) + dt = f (0) + f (∞) − f (0). (1.40)
s→0 0 dt
Therefore,
lim sF (s) = f (∞) = lim f (t) (1.41)
s→0 t→∞
which is known as the final value theorem. It is valid if all poles of sF (s)
are in the left half-plane.
From the definition of the Laplace transform, one can find the following
useful relation
Z ∞ Z ∞
f (t) dt = lim e−st f (t) dt = F (0) (1.42)
0 s→0 0
The Laplace transform and the inverse Laplace transform are

implemented in Wolfram Language as LaplaceTransform[f, t, s] and
InverseLaplaceTransform[F, s, t], respectively.
Example 1.1. Show that

√
√
1 a2
L √ e− 4t = πs−1/2 e−|a| s . (1.43)
t
From the definition of the Laplace transform, one has
Z ∞ Z ∞
1 − a2 1 a2 a2 2
L √ e 4t = √ e− 4t −st dt = 2 e− 4u2 −su du. (1.44)
t 0 t t=u2
0
Then, we consider the integral

Z ∞ 2
2
u2 − u
r
I(q, r) = e−q 2
du. (1.45)
0
Differentiating with respect to the parameter r, one obtains
Z ∞
∂ 2 2 r 2 du
I(q, r) = −2r e−q u − u2 2 . (1.46)
∂r 0 u
Using the variable change v = rq u, one finds

Z ∞
∂ 2 2 r2
I(q, r) = −2q e−q v − v2 dv = −2qI(q, r). (1.47)
∂r 0
The solution of this differential equation is
I(q, r) = C e−2qr , (1.48)
where the constant C is obtained from the condition r = 0. Thus,
Z ∞ √
−q 2 u2 π
I(q, 0) = e du = , (1.49)
0 2q
√
π
from where it follows that C = 2q .
Therefore, the integral becomes
√
π −2qr
I(q, r) = e . (1.50)
2q
√ a
Using that q = s and r = 2 we arrive at the result (1.43).
Example 1.2. Show that

h √ i b b2
L−1 e−b s = √ e− 4t , (1.51)
4πt3
where b > 0.
Differentiating with respect to the parameter b and using the result
obtained in the previous example, we have
h √ i h √ √ i
L−1 e−b s = L−1 s−1/2 s e−b s
∂ h √ i
= − L−1 s−1/2 e−b s
∂b
∂ 1 − b2 b b2
=− √ e 4t =√ e− 4t .
∂b πt 4πt3
1.1.3 Mellin transform

The Mellin transform plays an important role in the definition of the Fox
H-function. It is also a convenient technique for treating dilation opera-
d n
tors like [x dx ] . The transformation is described by the kernel function
K(q, x) = xq−1 , while the limits are set to a = 0 and b = ∞, see Table 1.1,
Z ∞
M[f (x)](q) = F (q) = f¯(q) = f (x) xq−1 dx. (1.52)
0
The Mellin transform function (the Mellin image) F (q) is defined on a
complex plane with complex values q = q1 + ıq2 such that the real part,
q1 is defined by f (x). For example, if f (x) = e−rx , Eq. (1.52) defines a
gamma function
Z ∞
Γ(q) = xq−1 e−x dx, (1.53)
0
where q1 > 0 (see Sec. 1.3.1 for details). In other example, for f (x) =
(x − x0 )z+ ≡ Θ(x)(x − x0 )z , the Mellin transform
M [Θ(x)(x − x0 )z ] (q) = x0z+q /(z + q) (1.54)
exists for q1 < −<(z). In general case, a region of all valid values of <(q)
is known as a strip of definition of the Mellin transform, see e.g., Ref. [1].
Another important example for f (x) = (1 + x)−1 defines a beta function:
Z ∞ q−1
x
F (q) = = B(q, 1 − q) , (1.55)
0 1+x
see Examples 1.5 and 1.6 of Sec. 1.3.1 for details.
A relation to the Laplace and Fourier transforms can be established by
change of the variables x = e−y and dx = −e−y dy. The integral (1.52)
reads
Z ∞
F (q) = g(y)e−qy dy , (1.56)
−∞
where g(y) ≡ f (e−y ). This expression corresponds to the two-sided Laplace
transform. To obtain the Fourier transform, one uses q = q1 + ıq2 ≡ a − ık
in Eq. (1.56), which according to definition (1.4) yields
Z ∞
F (k) = F g(y)e−ay (k) ≡ f e−y e−ay e−ıky dy.

(1.57)
−∞
Expression (1.57) is used to define the inverse Mellin transform in the
form of the inverse Fourier transform. That is
Z ∞
−y
−ay 1
f e e = F (k)eıky dk. (1.58)
2π −∞
Returning to the x variable, we have from Eq. (1.58) that the inverse Mellin
transform is given by
Z c+ı∞
1
M−1 [F (q)](x) = f (x) = F (q) x−q dq. (1.59)
2πı c−ı∞
It also follows from the inverse transform in Eq. (1.59), see [1], that if in the
strip of definition, F (q) is an analytic function and satisfies the inequality
|F (q)| < A|q|−2 for some constant A, then the function f (x) is a continuous
function of x ∈ [0, ∞) and its Mellin transform is F (q).
Additionally to the basic property of linearity see e.g., [1, 2], the Mellin
transform has the following properties:
(1) Multiplication by xa :
M [xa f (x)] = F (q + a) (1.60)
We easily show that

Z ∞
M [xa f (x)] = xa+q−1 f (x) dx = F (x + a).
0
(2) Transform of f (xa ):
1
M [f (xa )] = F (q/a) (1.61)
a
This property can be easily shown,
Z ∞
a
M [f (x )] = xq−1 f (xa ) dx
0
Z ∞
1
=a z (q−1)/a f (z) z (1−a)/a dz
(z=x ) 0 a
1 ∞ q/a−1
Z
1
= z f (z) dz = F (q/a).
a 0 a
(3) Scaling property (a > 0):
M[f (ax)] = a−q F (q) (1.62)
From the definition of the Mellin transform, we have

Z ∞
M[f (ax)] = xq−1 f (ax) dx
0
Z ∞
1
= z q−1 f (z) dz = a−q F (q).
(z=ax) aq 0
(4) Transform of convolution:

Z ∞
dr
M f (r) g (x/r) = F (q)G(q) (1.63)
0 r
(5) Transform of derivative of f (x):

n
d Γ(q)
M f (x) = (−1)n F (q − n) (1.64)
dxn Γ(q − n)
(6) Transform of logn x f (x):
dn
M [logn x f (x)] = F (q) (1.65)
dq n
For the latter expression, we have
Z ∞
n
M [log x f (x)] = xq−1 logn x f (x) dx
0
Z ∞
= z n eqz f (ez ) dz
(z=log x) −∞
Z ∞
∂n
= n eqz f (ez ) dz
∂q −∞
Z ∞
∂n dn
=z n
xq−1 f (x) dx = n F (q)
(x=e ) ∂q 0 dq
The Mellin transform of the Dirac δ-function yields
Z ∞
M [δ(x − x0 )] = δ(x − x0 ) xq−1 dx = x0q−1 . (1.66)
0
Another useful formula reads Z ∞
1 log2 x 1 log2 x
M √ e− 4α = √ e− 4α xq−1 dx
4πα 4πα 0
Z ∞ h 2
1
i
− z −qz
= √ e 4α dz
z=log x 2 4πα −∞
2 Z ∞ h √ i2
eαq − 2√ z
− αq 2
= √ e α
dz = eαq . (1.67)
2 4πα −∞
In Wolfram Language, Mellin and inverse Mellin transforms are imple-
mented as MellinTransform[f, x, q] and InverseMellinTransform[F, q, x],
respectively.
1.2 Asymptotic expansions
1.2.1 Tauberian theorems

The asymptotic behavior of a given function r(t) can be analyzed by means
of the Tauberian theorems [3]. One of the theorems states that if the
asymptotic behavior of r(t) for t → ∞ is given by
r(t) ' t−α , t → ∞, α > 0, (1.68)
then, the corresponding Laplace pair r̂(s) = L[r(t)] has the following be-
havior for s → 0
r̂(s) ' Γ(1 − α)sα−1 , s → 0. (1.69)
The theorem also works in the opposite direction, ensuring that r(t) is the
non-negative and monotone function at infinity.
This theorem can be formulated in the form of the so-called Hardy-
Littlewood theorem. The theorem states that, if the Laplace-Stieltjes
transform of a given non-decreasing function F with F (0) = 0, defined by
Stieltjes integral
Z ∞
ω(s) = e−st dF (t), (1.70)
0
has asymptotic behavior
ω(s) ' Cs−ν , s → ∞ (s → 0), (1.71)
where ν ≥ 0 and C are real numbers, then the function F has asymptotic
behavior
C
F (t) ' tν , t→0 (t → ∞). (1.72)
Γ(ν + 1)
These Tauberian theorems are widely used in the theory of anomalous dif-
fusion and in the theory of non-exponential relaxation processes.
The Tauberian theorem for slowly varying functions has also many ap-
plications in the theory of ultraslow diffusive processes and for analysis of
strong anomaly. The theorem states that if a function r(t), t ≥ 0, has the
Laplace transform r̂(s) whose asymptotics behaves as follows

1
r̂(s) ' s−ρ L , s → 0, ρ > 0, (1.73)
s
then
1 ρ−1
r(t) = L−1 [r̂(s)] ' t L(t), t → ∞. (1.74)
Γ(ρ)
Here L(t) is a slowly varying function at infinity, i.e.,
L(at)
lim = 1,
t→∞ L(t)
for any a > 0. The theorem is also valid if s and t are interchanged, that is
s → ∞ and t → 0.
1.2.2 Generating function formalism

The generating function formalism or z-transform method1 is often used
for Markov chains, see e.g. Ref. [5]. By means of the generating function
formalism, a probability function f (t) = fn , defined at discrete times t =
n, is transformed into its continuous counterpart, known as a generating
function,
X∞
F (z) = fn z n . (1.75)
n=0
Here z is a complex variable, such that |z| < R, where R is the radius
of convergence. In our considerations, R = 1. The advantage of the z-
transform (1.75) lies in the reduction of convolutions, which involve the
sequence {fn }, as for example in Markov chains, to algebraic expressions
for F (z). The latter can be easily solved and explicit expressions for the
generating function F (z) can be obtained. Then the sequence {fn } can be
obtained by the inverse z-transform,
I
1
fn = F (z)z −(n+1) dz. (1.76)
2πı c
Here C is a counterclockwise closed path encircling the origin and entirely in
the region of convergence. Equation (1.76) can be obtained from Eq. (1.75)
by multiplying both sides by z −(n+1) and integrating with respect to z. The
lhs of Eq. (1.75) yields the rhs of Eq. (1.76), while for the rhs of Eq. (1.75)
we obtain
∞ I ∞ Z 2π
1 X −(n−m+1) 1 X 1
fm z dz = fm eı(m−n)θ dθ
2πı m=0 c 2π m=0 2π 0
∞
X
= fm δm,n = fn . (1.77)
m=0
Here we take the contour as a circle with the radius R = 1 and make the
change of the variable z = eiθ .
In many cases, the knowledge of the explicit form of F (z) does not help
to calculate the integral, which can be very complicated. In that case, the
discrete Tauberian theorem for power series [3] is useful, as it establishes a
relation between fn and F (z).
Theorem 1.1 (The discrete Tauberian theorem [3]). Let fn ≥ 0 and
suppose that the power series
∞
X
F (z) = fn z n (1.78)
n=0
1 We follow Sec. 3.2 of Ref. [4] and Sec. XIII.5 of Ref. [3].
converges for 0 ≤ z < 1 and that

1 1
F (z) ∼ L , z → 1− , (1.79)
(1 − z)ρ 1−z
where L is a slowly varying function near infinity and 0 ≤ ρ < ∞. Then
Eq. (1.79) is equivalent to the following two relations,
1
f0 + f1 + · · · + fn−1 ∼ nρ L(n), n → ∞, (1.80)
Γ(ρ + 1)
and
1 ρ−1
fn ∼ n L(n), n → ∞. (1.81)
Γ(ρ)
The latter expression also supposes that {fn } is a monotonic sequence.
1.2.3 Laplace method

The Laplace method is used to estimate integrals of the form
Z b
I= e−tf (z) g(z)dz (1.82)
a
for large t. It can be approximated by2

s
−t f (z0 ) 2π
I≈e g(z0 ) , (1.83)
t|f 00 (z0 )|
where z0 is the extremum point of the function f (z), i.e., f 0 (z0 ) = 0, and
if the extremum point is within the integration limits (a < z0 < b). If
the extremum point is outside the integration limits (z0 > b), then the
approximation result is calculated at z0 = b.
Example 1.3. Let us evaluate the following integral

Z t
b2 y 2

ab
√ exp −at0 − 0 dt0
0 πt0 t
for large t and a > 0, b > 0.
2 There is a vast library on the issue. We just mention Refs. [6, 7].
We introduce t0 = tz, i.e., dt0 = t dz in the integral to obtain

Z 1
b2 y 2

ab
I= √ exp −atz − tdz
0 πtz tz
√ Z
ab t 1 b2 y 2

dz
= exp −t az + 2 √
π 0 t z z
√ Z 1
ab t
= √ e−tf (z) g(z)dz, (1.84)
π 0
where
b2 y 2 1
f (z) = az + and g(z) = √ . (1.85)
t2 z z
For the extremum point we find
b2 y 2 b|y|
f 0 (z)|z=z0 = a − =0 → z0 = √ . (1.86)
t2 z02 t a
This yields
√ √
b|y| b2 y 2 t a 2b a|y|
f (z0 ) = a √ + 2 = . (1.87)
t a t b|y| t
2b2 y 2 2b2 y 2 t3 a3/2 2a3/2 t

f 00 (z)|z=z0 = = = (1.88)
t2 z03 t2 b3 |y|y 2 b|y|
and
s
1/4 t
g(z0 ) = a . (1.89)
b|y|
If z0 < 1, then we apply the approximation formula (1.83) to obtain
√ s v
ab ta1/4 t u
u 2π −t
√
2b a|y|
I≈ √ e t , (1.90)
π b|y| t 2a3/2 t
t
b|y|
which gives
√ √
I ≈ b a e−2b a|y| . (1.91)
This result will be used later in the analysis of the transition to the
stationary state of a diffusing particle under stochastic resetting.
1.3 Special functions
1.3.1 Gamma function

The gamma function is defined for complex numbers with a positive real
part by the integral
Z ∞
Γ(α) = xα−1 e−x dx, α > 0. (1.92)
0
It is extended by analytical continuation to all complex numbers, except
the non-positive integers, where it has simple poles, see below. Integrating
by part, we obtain a recursion formula,
Γ(α + 1) = αΓ(α). (1.93)
An alternative representation of the gamma function is via Euler’s limit,
j!j α

Γ(α) = lim , (1.94)
j→∞ α(α + 1) . . . (α + j)
where Γ(1) = 1. For the integer α = n, the recursion (1.93) yields Γ(n+1) =
nΓ(n) = n!. Recursion (1.93) also defines the gamma function for negative
arguments as an analytical continuation Γ(z − 1) = Γ(z)/(z − 1), where
Γ(0) and Γ(−n) diverge, while the ratio is finite Γ(−n)/Γ(−m) = m!/n!.
The Gamma function is implemented in Wolfram Language by Gamma[z].
Example 1.4. Let us calculate Γ(1/2) = (−1/2)!:

Z ∞ Z ∞
2 √
Γ(1/2) = x−1/2 e−x dx = 2 e−y dy = π. (1.95)
0 0
Another useful function is the beta function with the integral represen-
tation for p, q > 0,
Z 1
B(p, q) = xp−1 (1 − x)q−1 dx, (1.96)
0
while for arbitrary p and q, it is defined as a composition of gamma func-
tions,
Γ(p)Γ(q)
B(p, q) = . (1.97)
Γ(p + q)
The beta function is implemented in Wolfram Language by Beta[a, b].
Using the beta function, some additional properties of the gamma func-
tion can be established as well, as shown in the ensuing examples.
Example 1.5. To prove formula (1.97) [8], we substitute x = y/(1+y)

in the definition (1.96) and obtain another integral representation of
the beta function,
Z ∞
B(p, q) = y p−1 (1 + y)−p−q dy. (1.98)
0
Next we establish the following identity
Z ∞
Γ(p + q)
xp+q−1 e−(1+y)x dx = . (1.99)
0 (1 + y)p+q
Carrying out the variable substitution z = (1 + y)x, we find
Z ∞ Z ∞ p+q−1
z dz
xp+q−1 e−(1+y)x dx = e−z
0 0 1+y 1+y
Z ∞
1
= z p+q−1 e−z dz, (1.100)
(1 + y)p+q 0
which leads to the relation (1.99), using the definition (1.92). We
multiply Eq. (1.99) by y p−1 and integrate with respect to y. Changing
the order of integration, we find
Z ∞ Z ∞ Z ∞
Γ(p)
dx xp+q−1 e−x y p−1 e−xy dy = dx xp+q−1 e−x p
0 0 0 x
Z ∞
= Γ(p) xq−1 e−x dx
0
Z ∞
= Γ(p + q) y p−1 (1 + y)−p−q dy, (1.101)
0
where we have used the identity (1.99) to evaluate the second integral.
Taking into account definitions (1.92) and (1.98), we obtain Eq. (1.97).
Example 1.6. Consider both expressions (1.96) and (1.97) for q =

1 − p:
Z 1 p−1
x dx
Γ(p)Γ(1 − p) = B(p, 1 − p) = . (1.102)
0 1 − x 1 −x
The integral converges for 0 < p < 1. Making the change of the
variable y = x/(1 − x), which yields dx/(1 − x) = dy/(1 + y), we
obtain
∞
y p−1 dy
Z
Γ(p)Γ(1 − p) = . (1.103)
0 1+y
We perform the analytical continuation in the complex plane with
y → z = reıφ and with a branch cut along the real axis (0, ∞), resulting
in the contour integral
I p−1
z dz
I= . (1.104)
C 1+z
The contour of the integration C consists of four parts: (1) the inte-
gration over a circle with large radius r = R → ∞, (2) the integration
over a circle around z = 0 with small radius r = → 0, and (3) two
integrals, one over the upper edge of the cut along the real axis (, R)
and the second over the lower edge of the cut in the interval (R, ).
There is only a simple pole at z = eıπ , which yields I = −2πıeıπp ,
according to the residue theorem. We have
I = IR + I + I(,R) (φ = 0) + I(R,) (φ = 2π) = 2πıeπıp , (1.105)
and for the individual integrals,
IR=∞ = I=0 = 0, (1.106a)
I0,∞ = B(p, 1 − p), (1.106b)
2πıp
I∞,0 = −e B(p, 1 − p). (1.106c)
Taking all these results into account, we finally obtain Euler’s reflection
formula,
π
Γ(p)Γ(1 − p) = . (1.107)
sin(pπ)
√
Equation (1.107) implies that Γ(1/2) = π, see also Example 1.4.
Example 1.7. Another property that follows immediately from the

beta function (1.96) is the Legendre duplication formula. From the
integral (1.96), we have
Z 1
B(p, p) = [τ (1 − τ )]p−1 dτ. (1.108)
0
Performing the change of the variable s = 4τ (1−τ ), which is symmetric

with respect to τ = 1/2, we find
Z 21 Z 1
1
B(p, p) = 2 [τ (1 − τ )]p−1 dτ = 2p−1 sp−1 (1 − s)−1/2 ds
0 2 0
1
= 2p−1 B(p, 1/2). (1.109)
2
√
Using the definition (1.97) and the fact that Γ(1/2) = π, we obtain
the Legendre duplication formula,
√
Γ(p)Γ(p + 1/2) = 21−2p π Γ(2p), 2p 6= −1. (1.110)
1.3.2 Contour integral representation of Γ(p) and 1/Γ(p)

We perform the analytical continuation in the complex plane z = x − iy in
the integration (1.92) to represent it as a contour integral
Z Z
p−1 −z
z e dz = e(p−1)[ln(z)]−z dz, (1.111)
C C
where the contour C starts at +∞, runs around the point z = 0, and ends
again at +∞. Since z = 0 is a branch point, we take a branch cut along
the non-negative real axis (0, ∞). In this case, the contour consists of three
parts: (1) the upper edge of the branch cut (∞, ), (2) the circle C of
radius → 0 with the center at z = 0, and (3) the lower edge of the branch
cut (, ∞). On the upper edge, ln z = ln x is real, while on the lower edge
ln z = ln x + 2πı. Therefore,
Z Z
z p−1 e−z dz = xp−1 e−x dx
C ∞
Z Z ∞
+ z p−1 e−z dz + e2(p−1)πı xp−1 e−x dx. (1.112)
C
In the limit → 0, integration over the contour C vanishes, and we obtain

Z ∞ Z
1
Γ(p) = xp−1 e−x dx = 2(p−1)πı z p−1 e−z dz. (1.113)
0 e −1 C
To obtain a formula for the contour integral representation of 1/Γ(p),
we replace p by 1 − p in Eq. (1.113),
Z
z −p e−z dz = (e−2pπı − 1)Γ(1 − p). (1.114)
C
With the further substitution z = τ eπı = −τ , we invert the contour C with

respect to the y axis. In this case, the contour is known as a Hankel contour
{Ha}. The contour integral (1.114) reads
Z Z
−p −z −pπı
z e dz = −e τ −p eτ dτ. (1.115)
C {Ha}
Combining Eqs. (1.114) and (1.107), we find

Z
2πı
τ −p eτ dτ = 2ı sin(pπ)Γ(1 − p) = . (1.116)
{Ha} Γ(p)
This implies the following integral representation of the reciprocal gamma
function,3
Z
1 1
= τ −p eτ dτ. (1.117)
Γ(p) 2πı {Ha}
1.3.3 Mittag-Leffler functions

1.3.3.1 One parameter Mittag-Leffler function
The one parameter Mittag-Leffler (M-L) function is defined by [9]
∞
X zk
Eα (z) = , (1.118)
Γ(αk + 1)
k=0
where z ∈ C, <(α) > 0) and Γ(z) is the gamma function. It is an entire

function of order ρ = 1/<(α) and type 1.
It is a generalization of the exponential function
∞ ∞
X (±z)k X (±z)k
E1 (±z) = = = e±z . (1.119)
Γ(k + 1) k!
k=0 k=0
as well as trigonometric and hyperbolic functions

∞ ∞
X (−1)k z 2k X (−1)k z k
E2 (−z 2 ) = = = cos(z), (1.120)
Γ(2k + 1) (2k)!
k=0 k=0
∞ ∞
X z 2k X z 2k
E2 (z 2 ) = = = cosh(z). (1.121)
Γ(2k + 1) (2k)!
k=0 k=0
3 By deformation of the contour, this expression is the inverse Laplace transform.

Problem 1.1. Show that

" √ !#
1 z1/3 1/3 3 1/3
E3 (z) = e + 2e−z /2 cos z (1.122)
3 2
and
1h i
E4 (z) = cos z 1/4 + cosh z 1/4 . (1.123)
2
For α = 1/2 one has

∞
√ X (±1)k z k/2 √
= ez 1 + erf(± z ) ,

E1/2 (± z) = (1.124)
Γ(k/2 + 1)
k=0
where erf(z) is the error function defined by the series

∞
2 X (−1)k z 2k+1
erf(z) = √ , (1.125)
π k!(2k + 1)
k=0
and by the integral

Z z
2 2
erf(z) = √ e−x dx. (1.126)
π 0
Here we note that the result can be presented in terms of the complementary
error function which is defined by
Z ∞
2 2
erfc(z) = 1 − erf(z) = √ e−x dx, (1.127)
π z
which satisfies
erfc(−z) = 2 − erfc(z). (1.128)
Therefore, one has
√ √ √
E1/2 (± z) = ez 2 − erfc(± z ) = ez erfc(∓ z ).

(1.129)
The error function and the complementary error function are imple-
mented in Wolfram Language as Erf[z] and Erfc[z], respectively.
Example 1.8. Show that for n ∈ N the following relation holds true:
dn
En (±z n ) = ±En (±z n ) . (1.130)
dz n
By using the definition (1.118), one finds

∞ ∞ dn nk
dn n dn X (±1)k z nk X (±1)k dz nz
En (±z ) = n =
dz n dz Γ(nk + 1) Γ(nk + 1)
k=0 k=0
∞
X nk(nk − 1) . . . (nk − n)(±1)k z nk−n
=
Γ(nk + 1)
k=1
∞
X (±1)k z n(k−1)
=
Γ(n(k − 1) + 1)
k=1
∞
X (±1)k z nk
=± = ±En (±z n ) .
Γ(nk + 1)
k=0
The integral representation of the one parameter M-L function is

Z α−1 s
1 s e
Eα (z) = α
ds. (1.131)
2πı C s − z
The path C is a loop starting and ending at −∞ and encircles |s| ≤ |z|1/α
in the positive sense −π ≤ arg s ≤ π on C. From the integral representation
one can analyze the asymptotic behaviors of the M-L function.
For 0 < α < 2, the one parameter M-L function has the following
asymptotics [10, 11]
n
1 z1/α X z −k
+ O |z|−1−n ,

Eα (z) = e − |z| → ∞, | arg z| ≤ θ,
α Γ(1 − αk)
k=1
(1.132)
n
X z −k
+ O |z|−1−n ,

Eα (z) = − |z| → ∞, θ ≤ | arg z| ≤ π,
Γ(1 − αk)
k=1
(1.133)
where πα/2 < θ < min{π, απ}.
Therefore, we can use the following asymptotics of the one parameter
M-L function
1 1/α
Eα (z) ∼ ez , z 1, (1.134)
α
for 0 < α < 2. Moreover, we will use the formula [10, 11, 12],
∞ ∞
X (−z)−n−1 X (−z)−n
Eα (−z) = − =− , z > 1. (1.135)
n=0
Γ(1 − α(1 + n)) n=1
Γ(1 − αn)
for 0 < α < 2, to find the asymptotic behavior of the one parameter M-L
function Eα (−z). For large z we have
z −1 z −2
Eα (−z) ∼ − , z 1. (1.136)
Γ(1 − α) Γ(1 − 2α)
The most important one parameter M-L function in the theory of frac-
tional differential equations is the associated one parameter M-L function
Eα (t; λ) = Eα (−λtα ) (α > 0; λ ∈ C) . (1.137)
Its Laplace transform reads [13]
sα−1
L [Eα (t; ∓λ)] = , (1.138)
sα ∓ λ
where <(s) > |λ|1/α .
From the asymptotic behaviors of the one parameter M-L function, one
has
1 1
Eα (t; −λ) = Eα (λtα ) ∼ exp [λtα ]1/α = exp λ1/α t , t 1,
α α
(1.139)
and
t−α t−2α
Eα (t; λ) = Eα (−λtα ) ∼ − 2 , t 1, (1.140)
λΓ(1 − α) λ Γ(1 − 2α)
for large t.
For t 1, the one parameter M-L function behaves as follows
λtα λtα

Eα (t; ∓λ) = Eα (±λtα ) ∼ 1 ± ' exp ± , t 1.
Γ(α + 1) Γ(α + 1)
(1.141)
For 0 < α < 1 it is a stretched exponential, while for 1 < α < 2 it is a
compressed exponential.
Example 1.9. Find the first term of the asymptotic behavior of the
associated one parameter M-L function for t 1, given by Eα (t; λ) =
t−α
Eα (−λtα ) ∼ λΓ(1−α) (λ > 0), by using the Tauberian theorem.
Let us use the Laplace transform formula (1.138) for the associated
α−1
one parameter M-L function L [Eα (t; λ)] (s) = ssα +λ . By its series
expansion for s → 0 we have
1 sα−1 sα sα−1

1 α−1
L [Eα (t; λ)] = ∼
sα s→0 s 1 − ∼ .
λ1+ λ λ λ λ
By using the Laplace transform of power function (1.34) and applying
the Tauberian theorem, we find the asymptotic behavior for t → ∞,

1 t−α
α−1
s
Eα (t; λ) ∼ L−1 = , α < 1.
t→∞ λ λ Γ(1 − α)
Example 1.10. Find the second term in the asymptotic expansion of

the associated one parameter M-L function for t 1, by using the
Tauberian theorem.
We use the same approach as in Example 1.9. Therefore, we have
1 sα−1 sα sα−1 s2α−1

1 α−1
L [Eα (t; λ)] = ∼
sα s→0 s 1 − ∼ − .
λ1+ λ λ λ λ λ2
Since 0 < α < 1 we note that β one can not directly use the Laplace
Γ(β+1)
transform formula (1.34), L t = sβ+1 , β > −1, for the second
term. For this reason, we first divide the both sides of the equation by
s, to have
1 sα−2 s2α−2
L [Eα (t; λ)] ∼ − ,
s s→0 λ λ2
and then we find the inverse Laplace transform. Thus, we find
Z t α−2
s2α−2 t1−α t1−2α

s
Eα (t0 ; λ) dt0 ∼ L−1 − 2
= − 2 .
0 t→∞ λ λ λΓ(2 − α) λ Γ(2 − 2α)
By differentiating both sides we finally find
t−α t−2α
Eα (t; λ) ∼ − 2 .
t→∞ λΓ(1 − α) λ Γ(1 − 2α)
The one parameter M-L function (1.118) is implemented in the Wolfram

Language as MittagLefflerE[α, z].
1.3.3.2 Two parameter Mittag-Leffler function

The two parameter M-L function is defined by [14, 15, 16, 17]
∞
X zk
Eα,β (z) = , (1.142)
Γ(αk + β)
k=0
with z, β ∈ C, <(α) > 0. It is an entire functions of order ρ = 1/<(α) and
type 1. The basic properties and relations of the one and two parameter
Another random document with
no related content on Scribd:
That some kind of message passed between the two dogs is, I think, beyond a
reasonable doubt; and it is precisely this silent and mysterious kind of
communication (the kind that occurs when your [9]dog comes to you when you
are reading, looks intently into your face, and tells you without words that he
wants a drink or that it is time for him to be put to bed) that I propose now to
make clear. Before we enter that trail of silence, however, there is a much
simpler language, such as is implied in the whistle of a quail or the howl of a
wolf, which we must try as best we can to interpret. For unless our ears are
keen enough to distinguish between the food and hunting calls of an animal, or
between bob-white’s love note and the yodel that brings his scattered flock
together, it will be idle for us to ask what message or impulse a mother wolf
sends after a running cub when she lifts her head to look at him steadily, and
he checks his rush to return to her side as if she had made the murky woods
echo to her assembly clamor.
[10]
[Contents]
II
Cries of the Day and Night
The simplest or most obvious method of animal communication is by

inarticulate cries, expressive of hunger, loneliness, anger, pleasure, and other
primal needs or emotions. The wild creatures are mostly silent, and so is the
bulk of their “talk,” I think; but they frequently raise their voice in the morning or
evening twilight, and by observing them attentively at such a time you may
measure the effect of their so-called language. Thus, you see plainly that to
one call the animal cocks his ear and gives answer; at another call he
becomes wildly excited; a third passes over him without visible result; a fourth
sets his feet in motion toward the sound or else [11]sends him flying away from
it, according to its message or import.
That animal cries have a meaning is, therefore, beyond serious doubt; but
whether they have, like our simplest words, any definite or unchanging value is
still a question, the probable answer being “No,” since a word is the symbol of
a thought or an idea; but animals live in a world of emotion, and even our
human emotions are mostly dumb or inarticulate. I must give this negative
answer, notwithstanding the fact that I have learned to call various birds and
beasts, and that I can meet Hotspur’s challenge on hearing Glendower boast
that he can call spirits from the vasty deep:
Why, so can I, or so can any man;

But will they come when you do call for them?
Yes, the birds and beasts will surely come if you know how to give the right
call; but I am still doubtful whether among themselves their audible cries are
ever quite so intelligible as is their silence.
This question of animal speech has received a different and more positive
answer, by the way, from a man who has spent many years in persistent
observation of wild apes and monkeys. After watching the lively creatures from
his cage in the jungle, attracting them by means of various fruits and recording
their jabber in a phonograph, he [12]claims to have discovered the monkey
words for food, water, danger and other elementary matters. Moreover, when
his phonograph repeats these simian words the monkeys of another locality
seem to understand them, since they run to the proper dish at the word “food”
or show evident signs of alarm at the word “danger.”
It is doubtless much easier to deny such a conclusion than to prove or

disprove it; but denial is commonly the first refuge of ignorance and the last of
dogmatism, and with these we are not concerned. I do not know whether
Garner claims too much or too little for his monkeys; I have never had
opportunity to test the matter in the jungle, and the caged monkeys with which
I have occasionally experimented are too debased of habit or too imbecile in
their affections to interest one who has long dealt with clean wild brutes. At
times, however, when I have watched a monkey with an organ-grinder, I have
noticed that the unhappy little beast displays a lively interest in the chitter of
chimney-swifts—a lingo which to my dull ears sounds remarkably like monkey-
talk. But that is a mere impression, momentary and of little value; while Garner
speaks soberly after long and immensely patient observation.
To return to first-hand evidence: among wild creatures of my acquaintance the

crows come [13]nearer than any others to something remotely akin to human
speech. Several times I have known a tame crow to learn a few of our words
and, what is much more significant, to show his superiority over parrots and
other mere mimics by using one or more of the words intelligently. There was
one crow, for example, that would repeat the word “hungry” in guttural fashion
whenever he thought it was time for him to dine. He used this word very
frequently when his dinner or supper hour drew nigh, giving me the
impression, since he did not confuse it with two other words of his vocabulary,
that he associated the word with the notion of food or of eating; and if this
impression be true to fact, it indicates more than appears on the surface. We
shall come to the wild crows and their “talk” presently; the point here is, that if
this bird could use a new human word in association with a primal need, there
is nothing to prevent him from using a sound or symbol of his own in the same
way. In other words, he must have some small faculty of language.
Another tame crow, which an imaginative boy named Pharaoh Necho because
of his hippety-hop walk, proved himself inordinately fond of games, play, social
gatherings of every kind. To excitement from any source, whether bird or brute
or human, he was as responsive as a weather-vane; [14]but his play ran mostly
to mischief, or to something that looked like joking, since he could never see a
contemplative cat or a litter of sleepy little pigs without going out of his way to
tweak a tail and stir up trouble. At times he would watch, keeping out of sight in
a leafy tree or on the roof of the veranda, till Tabby, the house cat, came out
and sat looking over the yard, her tail stretched out behind her. If she lay down
to sleep, or sat with tail curled snugly around her forepaws, she was never
molested; but the moment her tail was out of her sight and mind Necho saw
the chance for which he had apparently been waiting. Gliding noiselessly down
behind the unconscious cat he would tiptoe up and hammer the projecting tail
with his beak. It was a startling blow, and at the loud squall or spitting jump that
followed he would fly off, “chuckling” immoderately.
When Necho saw or heard a gang of boys assembled he would neglect even
his dinner to join them; and presently, without ever having been taught, he
announced himself master of a new art by yelling, “Ya-hoo! Come on!” which
was the rallying-cry of the clan in that neighborhood. He said this in ludicrous
fashion, but unmistakably to those who knew him. Sometimes he would croak
the words softly to himself, as if memorizing them or pleased at the sound; but
for the most [15]part he waited till boys were gathering for a swim or a ball
game, when he would launch himself into flight and go skimming down the
road, whooping out his new cry exultantly. What meaning he attached to the
words, whether of boys or fun or mere excitement, I have no means of
knowing.
After learning this much of our speech Necho took to the wild, following a call
of the blood, I think; for it was springtime when he disappeared, and the crows’
mating clamor sounded from every woodland. These birds are said to kill every
member of their tribe who returns to them after living with men, and the saying
may have some truth in it. I have noticed that many tame crows are like tame
baboons in that they seem mortally afraid of their wild kinsmen; but Necho was
apparently an exception. If he had any trouble when first he returned to his
flock, the matter was settled without our knowledge, and during the following
autumn there was evidence that he was again in good standing. Long
afterward, as I roamed the woods, I might hear his lusty “Ya-hoo! Come on!”
from where he led a yelling rabble of crows to chivvy a sleeping owl or jeer at a
running fox; and occasionally his guttural cry sounded over the tree-tops when
I could not see him or know what mischief was afoot. He never [16]returned to
the house, and never again joined our play or allowed a boy to come near him.
Not all crows have this “gift of speech”; and the fact that one tame crow learns
to use a few English words, while five or six others hold fast to their own lingo,
has led to the curious belief that, if you want to make a crow talk, you must
split his tongue. How such a belief originated is a mystery; but it was so fixed
and so widespread when I was a boy that no sooner was a young crow taken
from a nest than jack-knives were sharpened, and the leathery end of the
crow’s tongue was solemnly split after grave debate whether a seventh or a
third part was the proper medicine. If the crow talked after that, it was proof
positive that the belief was true; and if he remained dumb, it was a sign that
there was something wrong in the splitting; which is characteristic of a large
part of our natural-history reasoning. The debates I have heard or read on the
“unanswerable” question of how a chipmunk digs a hole without leaving any
earth about the entrance (a question with the simplest kind of an answer) are
mostly suggestive of the split-tongue superstition of crow language.
Of the tame crows I have chanced to observe, only a small proportion showed
any tendency to repeat words; and these gifted ones are, I judge, [17]the same
crows that in a wild state may occasionally be heard whistling like a jay, or
“barking” or “hooting” or making some other call which ordinary crows do not or
cannot make, and which shows an individual talent of mimicry. This last, which
I have repeatedly observed among wild crows, is a very different matter from
speech; but from the fact that these mimics learn to use a few English words
more or less intelligently one might not be far wrong in concluding that every
crow has in his brain a small undeveloped nest of cells corresponding to our
“bump” of language.
A closer observation of the wild birds may confirm this possibility. Thus, when
you hear a solitary crow in a tree-top crying, “Haw! Haw!” monotonously,
dipping his head or flirting his tail every time he repeats it, you may be sure
that somewhere within range of his eye or voice a flock of his own kind are on
the ground, feeding. That this particular haw is a communication to his fellows,
telling them that the sentinel is on watch and all is well, seems to me very
probable. There are naturalists, I know, who ingeniously resolve the whole
phenomenon into blind chance or accident; but that does not square very well
with the intelligence of crow nature as I have observed it; nor does it explain
the fact that once, when I avoided the sentinel and crept near enough to
[18]shoot two members of the flock he was supposedly guarding, the rest were
no sooner out of danger than they whirled upon the recreant and beat him
savagely to the ground.
If you are interested enough to approach any crow-sentinel in a casual or

indifferent kind of way (he will take alarm if you approach quickly or directly),
you must note that his haw changes perceptibly while you are yet far off. It is
no longer formal or monotonous; nor is it uttered with the same bodily attitude,
as your eyes plainly see. You would pronounce and spell the cry exactly as
before (it should be written aw or haw, not caw, for there is no consonant
sound in it); but if your ears are keen, they will detect an entirely different
accent or inflection, as they detect different accents and meanings when a
sailor’s casual or vibrant “Sail ho!” sings down from the crow-nest of a ship.
Now run a few steps toward the sentinel, or pretend to hide and creep, and
instantly the haw changes again. This time the accent is sharper even to your
dull ears; and hardly is the cry uttered when all the crows of the unseen flock
whirl into sight, heading swiftly away to the woods and safety.
Apparently, therefore, this simple haw of the crow is like a root word of certain
ancient languages, the Chinese, for example, which has several [19]different
intonations to express different ideas, but which all sound alike to foreign ears,
and which are spelled alike when they appear in foreign print. To judge by the
crows’ action, it is certain that their elementary haw has at least three distinct
accents to express as many different meanings: one of “all’s well,” another of
“watch out,” and a third of “be off!” Moreover, the birds seem to understand
these different meanings as clearly as we understand plain English; they feed
quietly while haw means one thing, or spring aloft when it means another; and
though you watch them a lifetime you will see nothing to indicate that there is
any doubt or confusion in their minds as to the sentinel’s message.
Not only the crows, but the wild ducks as well, and the deer and the fox and
many other creatures, seem to understand crow-talk perfectly, or at least a part
of it which concerns their own welfare. Thus, on the seacoast in winter you
hear the crows hawing continually as they follow the tide-line in search of food.
For hours this talk goes on, loudly or sleepily, and the wild ducks pay
absolutely no attention to it; though they must know well that hungry crows will
kill a wounded or careless duck and eat him to the bones whenever they have
a chance. Because of this dangerous propensity you would naturally expect
the water-fowl [20]to be suspicious of the black freebooter and to be alert when
they see or hear him; but no sooner do you begin to hunt with a gun than you
learn a thing to make you respect the crow, and perhaps to make you wonder
how much or how very little you know of the ways of the wood folk.
Many of the ducks, the black or dusky mallards especially, like to come ashore
every day in a secluded spot under the lee of a bank, there to rest or preen or
take a quiet nap in company. It is a tempting sight to see a score or a hundred
of the splendid birds in a close group, their heads mostly tucked under their
wings; but it is practically impossible to stalk them, for the reason that the
crows are forever ranging the shore, and a crow never passes a group of
sleeping ducks without lifting his flight to take a look over the bank behind
them. What his motive is no man can say; we only note that, in effect, he
stands sentinel for the ducks against a common enemy, as he habitually does
for his own kind. There is no escaping that keen, searching glance of his; he
sees you creeping through the beach-grass or hiding behind a bush. He flings
out a single haw! with warning, danger, derision in it; and now the same ducks
that have heard him all day without concern spring aloft on the instant and
head swiftly out to sea.
He flings out a single “Haw!” and the ducks spring aloft on the
instant and head swiftly out to sea.
[21]
The crows have several other variations of the same cry, expressive of other
matters, which all the tribe seem to understand clearly, but which are
meaningless to human ears. When I imitate the distress-call of a young crow,
for example, I can bring a flock over my head at almost any time, the only
condition being that I keep well concealed. At the first glimpse of a man in
hiding they sheer off, and it is seldom that I can bring them back a second time
to the same spot; yet I have a companion, one who utters a call very much like
mine to ordinary ears, who can bring the flock back to him even after they have
seen him and suffered at his hands. More than once I have stood beside him
in the woods and fired a gun repeatedly, killing a crow and scattering the flock
pell-mell at every shot; but no sooner does he begin to talk crow-talk than back
they come again. What he says to them that I do not or cannot say is
something that only the crows understand.
It is commonly assumed that they come to such a call because they hear in it a
cry for help from one of their own kind. That is undoubtedly true at times; for a
help-call, especially from a cub or nestling, is a summons to which most
animals and birds instinctively respond. And, strangely enough, the smaller
they are the braver they seem to be. [22]A mother-partridge has more than
once flown in my face or beaten me with her wings, while “fierce” hawks, owls
and eagles have merely circled around me at a safe distance when I came
near their young. In the majority of cases, however, I think that birds come to a
distress-call simply because the excitement of an individual spreads to all
creatures within sight or hearing, just as a crowd of men or women will become
excited and rush to a common center before they know what the stir is all
about.
In confirmation of this theory, it is not necessary to cry like a distressed young

crow to bring a flock over your head. The imitated hawing of an old crow will do
quite as well, if you throw the proper excitement into it. Again, on any summer
day you will hear in your own yard the pip-pip of arriving or departing robins.
The same call is uttered by both sexes, at all times and in all places; yet if you
listen closely you must note that there is immense variety in the accent or
inflection of even this simple sound. The call is clear, ringing, joyous when the
robins first arrive in the spring; it is subdued when they gather for the autumn
flight; it is sleepy or querulous when they stand full-fed by the nest, and most
business-like when they launch themselves into flight, which is the moment
when you are most sure to hear it. [23]A robin utters this call hundreds of times
every day, in one accent or another, and neither the other robins nor their
feathered neighbors seem to pay any attention to it; but when a red squirrel
comes plundering a nest, and the mother robin sends forth the same pip-pip
with a different intonation, then the response is instantaneous. The alarm
spreads swiftly over wood and field; clamor uprises, and birds of many species
come rushing in from all directions; not because they have heard that Meeko is
again killing young robins (at least, it does not seem so to me), but because
excitement is afoot, and they are bound to join it or find out about it before they
can settle down comfortably to their own affairs.
There is an interesting way by which you may test this contagion of excitement
for yourself. Hide at the edge of the woods or in any other bird neighborhood in
the early morning, preferably at a season when every nest has eggs or
fledglings in it; press two fingers against your lips and draw the breath sharply
between them, repeating the squeaky cry as rapidly as possible. The sound
has a peculiarly exciting quality even to human ears (twice have I seen men
run wildly to answer it), and birds come to it as boys to a fire alarm. In a few
moments you may have them streaming in from the four quarters of bird world,
all highly [24]excited, and perhaps all ready to protect some innocent nest from
snake or crow or squirrel. Because the response is most electric at the season
when fledglings are most helpless, you are apt to think that this call of yours is
mistaken by mother birds for a cry for help. That may be true; but be not too
sure about it. The fledglings themselves will come almost as readily to the call
when the nesting season is over and gone.
I have tried that same exciting summons in many places, wild or settled, and
commonly but not invariably with the same result, as if it were a word from the
universal bird language. Once in a secluded valley of northern Italy I saw a
hunter with his gun, and promptly forgot my own errand in order to chum with
him and find out what he had learned of the wood folk. He was hunting birds to
eat. “Those birds there!” he said, pointing to a passing flock which I did not
recognize, but which seemed pitifully small game to me. Presently I learned
that he could not shoot flying, and was having such bad luck that, he said, the
devil surely had a hand in it. He was a smiling, companionable loafer, and for a
time I tagged after him, watching him amusedly as he made careful but vain
stalks of little birds that seemed to have been made wild by much hunting. In a
spirit of thoughtless curiosity, and perhaps also [25]to test bird nature in a
strange land, I invited the hunter to hide with me in a thicket while I gave the
call which had so often brought the feathered folk of my own New England
woods. At my cry a wisp of birds whirled in to light at the edge of the covert;
the Italian’s gun roared; and then I discovered that the wretch was killing
skylarks.
I have since had many an uncomfortable moment at the thought of how many
lovely songsters may have paid with their lives for that ungodly experiment; for
my companion hailed me as a master Nimrod from the New World; and when I
refused, on the plea of bad luck, to teach him the call, I heard him give a
distressingly good imitation of it. Yet the experiment seemed to prove that
everywhere birds quickly catch the contagion of excitement; that in many
cases they respond to a call because it stirs their anger or curiosity rather than
because it conveys any definite summons for help or warning of danger.
When you open your ears among the beasts you hear precisely the same
story; that is, certain cries apparently have definite meaning, like the accented
haw of a crow, while others convey and also spread a wild emotion. Of all
beasts, the wolves are perhaps the keenest, the most intelligent, and these
seem to have definite calls for food or help or hunting or assembly. Such calls
are [26]strictly tribal, I think, like the dialects of Indians, since the call of a
coyote is quite different from the call of a timber wolf even when both intend to
convey the same meaning. A friend of mine, an excellent mimic, who spent
many years in the West, has shot more than a score of coyotes after drawing
them within range by sending forth the food-call in winter; but though he knows
also the food-call of the timber wolf, he has never once deceived these larger
brutes by his imitation of it; nor has he ever seen a wolf of one species
respond to the food or hunting call of another.
Like most other wild animals, timid or savage, the sensitive wolves all respond,
but much more warily than the birds, to almost any inarticulate cry expressive
of emotional excitement; just as your dog, who is yesterday’s wolf, grows
uneasy when you whine in your nose like a distressed puppy, or leaps up,
ready to fly out of door or window, when a wild ki-yi breaks out in the distance.
Indeed, it is easier to keep a boy from a fire than a dog from a crowd or
excitement of any kind; and the same is true of their wild relatives, though the
wariness of the latter keeps them hidden where you cannot follow their action.
The greatest commotion I ever witnessed in a timber-wolf pack was
occasioned by the moaning howl of a wounded wolf on a frozen lake in
midwinter. [27]It was a cry utterly unlike anything I had ever recorded up to that
time, and every time they heard it the grim beasts ran wildly here and there,
howling like lunatics. Then, when the wounded one grew quiet, they would
approach and sniff him all over; after which some would sit on their tails and
watch him closely, while others circled about on the ice, using their noses like
hounds in search of a lost trail.
Occasionally, when I have had these uncanny brutes near me in the North, I
have tried to call them or make them answer by giving what seemed to me a
very good imitation of their cries; but seldom has a howl of mine been
returned. On the contrary, the brutes almost always stop their howling
whenever I begin to talk wolf-talk, as if they were listening and saying, “What
under the moon is that now?” Then old Tomah, the Indian, comes out of his
blanket and gives a howl exactly like mine, but with something in it which I
cannot fathom or master, and instantly from the snow-filled woods comes back
the wild wolf answer.
Likewise, I have called moose in many different localities, and am persuaded

that it makes very little difference what kind of whine or grunt or bellow you
utter, since anything resembling a moose-call will do the trick if you know how
to [28]put the proper feeling into your voice. After listening carefully to many
callers, I note this characteristic difference: that one man invariably makes the
game wary, suspicious, fearful, no matter how finely he calls; while another in
the same place, with the same trumpet and apparently with the same call,
manages to put something into his voice, something primal, emotional and
essentially animal, which brings a bull moose hurriedly to investigate. Thus it
happens that the worst caller I ever heard—worst in that he had no sense, no
cunning, no knowledge of moose habits, and uttered a blatant, monstrous roar
unlike anything a sane man ever heard in the heavens above or the earth
beneath—was still the most successful in getting his game into the open.
Three nights in succession I heard him call in a region where moose were
over-shy from much hunting, and where my own imitation of the animal’s
natural voice brought small response. In that time fourteen bulls answered him,
all that were within hearing, I think; and every one of the great brutes threw
caution to the dogs and came out on the jump.
From such observations, and from others which I have not chronicled, I judge
that the higher orders of birds and beasts have a few calls which stand for
definite things, or mental images of [29]things, but that their ordinary cries
merely project an emotion or excitement in such a way that it stirs a similar
emotion in other birds or beasts of the same species; just as the sound of
hearty laughter invariably stirs the feeling of mirth in men who hear it, or any
inarticulate cry of fear sets human feet in motion—toward the cry if the hearer
be brave, or away from it if he be of cowardly disposition. Yet even among
men, who by civilization have lost some of their natural virtues, the primal
impulse still lives. Like the wolf or the raccoon, the man’s first impulse is to
rush to his distressed or excited fellows. If he turns and runs the other way, it
means simply that his artificial habit or training has deadened his natural
instincts.
In speaking of “man” here I refer to the genus homo, not to the male specimen
thereof. Among brutes most of the natural instincts are the same in both sexes;
they vary in degree, not in kind, and the instincts of the female are commonly
the stronger or keener. Yet I have noticed, or think I have noticed, this
difference: when a cry of distress is uttered in the woods, the first bird or beast
to appear is almost always a female; but the male is quicker on his toes at a
battle-yell or a senseless clamor.
This last is a personal impression, and cannot well be verified. The only record
I have which [30]might pass for evidence in the matter comes from my
observation of the crows. In the spring many of these questionable birds
indulge their taste for eggs or tender flesh and soon become incurable nest-
robbers; and for that reason I often shoot them, to save other and more useful
birds. The method is very simple: one hides and calls, and takes the crows as
they appear in swift flight, the number shot being commonly limited to one or
two at a time. And I have observed repeatedly, at different times and in
different localities, that when I use the distress-call of a young crow as a decoy,
the first to appear over the tree-tops is a female. This is the common rule, with
occasional exceptions to point or emphasize it. But whenever I clamor like a
crow that has discovered an owl, or send forth a senselessly excited hawing,
almost invariably the first crow to come whooping over is a long-winged and
glossy old male.
Does it seem to you like thoughtless barbarity on my part to kill crows in this
fashion? Perhaps it is barbarous; I do not quite know; but it certainly is not
thoughtless. One cannot blame the crows for their taste in eggs or nestlings;
but one must note that they destroy an enormous number of insectivorous
birds, and that the harm they do in this respect outweighs their usefulness in
destroying field-mice and beetles. I write this [31]with regret; for I admire the
crow, and consider him as, of all birds, the most intelligent and the most
considerate of his own kind. I know that it is a moot question whether the crow
does more harm or good, and that some naturalists have settled it in his favor;
but I have too often caught him plundering nests in the springtime to be much
impressed by his alleged usefulness at other seasons. I think that he may have
been once useful in preserving the so-called balance of nature; but that
balance is now dangerously unequal. The crow has flourished even in well-
settled regions, thanks to his superior wit, while other useful birds have
fearfully diminished, and this at a time when our orchards and gardens call
more and more insistently for their help. Because of his disproportionate
numbers the crow now appears to me, like our destructive and useless cats,
as a positive menace in a country where he once occupied a modest or
inconspicuous place—such a place as he still occupies in the wilderness,
where I meet him but rarely, and where I am glad to leave him in peace, since
he does not seriously interfere with his more beautiful or more useful
neighbors. But we are wandering from the dim trail of animal communication,
which we set out to follow.
The inarticulate but variously accented cries of which we have spoken

constitute the only animal [32]language to which our naturalists have thus far
paid any attention; and doubtless some of them would object to the use of the
word “language” in such a connection. In all matters of real natural history,
however (real, that is, in the sense of dealing at first hand with individual birds
or beasts), I am much more inclined to listen to old Tomah, who says, when I
ask him whether animals can talk: “Talk? Course he kin talk! Eve’ting talk in
hees own way. Hear me now make-um dat young owl talk.” And, stepping
outside the circle of camp-fire light, Tomah utters a hoot, which is answered at
a distance every time he tries it. After parleying with the stranger in this
tentative fashion, Tomah sends forth a different call; and immediately, as if in
ready acceptance of an invitation, a barred owl glides like a gray shadow into a
tree over our heads. I have heard that same old Indian use horned-owl talk,
wolf and beaver and woodpecker talk, and several other dialects of the wood
folk, in the same fascinating and convincing way.
One must judge, therefore, that most cries of the day or night have their
meaning, if only one knows how to hear them; yet they constitute but a part,
and probably a very small part, of the animal’s habitual communication with his
fellows. The bulk of it appears to be of that silent kind which [33]passed
between Don and Nip, and which, I have reason to believe, is the common
language of the whole animal kingdom.
To prove such a matter is plainly impossible. Even to investigate it frankly is to

enter a shadowy realm between the conscious and subconscious states,
where no process can be precisely followed, and where the liability to error is
always present. Let us therefore begin on familiar ground by examining certain
phenomena which we cannot explain, to be sure, but which have been
observed frequently enough to give us confidence that we are dealing with
realities. I refer especially to that curious warning or “feeling” of impending
danger, which is supposed (erroneously, I think) to depend upon the so-called
sixth sense of animals and men.
[34]
[Contents]
III
Chumfo, the Super-sense
For the word chumfo I am indebted to a tribe of savages living near Lake
Mweru, in Africa, and am grateful to them not only for naming a thing which
has no name in any civilized language, but also for an explanation of its
function in the animal economy. We shall come to the definition of the word
presently, after we have some clear notion of the thing for which the word
stands. As Thomas à Kempis says, if I remember correctly, “It is better to feel
compassion than to know how to define it.”
By way of approach to our subject, let it be understood that chumfo refers in a

general way to the animal’s extraordinary powers of sense perception,
[35]which I would call his “sensibility” had not our novelists bedeviled that good
word by making it the symbol of a false or artificial emotionalism. Every wild
creature is finely “sensible” in the true meaning of the word, his sensitiveness
being due to the fact that there is nothing dead or even asleep in nature; the
natural animal or the natural man is from head to foot wholly alive and awake.
And this because every atom of him, or every cell, as a biologist might insist, is
of itself sentient and has the faculty of perception. Not till you understand that
first principle of chumfo will your natural history be more than a dry husk, a
thing of books or museums or stuffed skins or Latin names, from which all
living interest has departed.
I am sometimes asked, “What is the most interesting thing you find in the
woods?” the question calling, no doubt, for the name of some bird or beast or
animal habit that may challenge our ignorance or stir our wonder. The answer
is, that whether you search the wood or the city or the universe, the only
interesting thing you will ever find anywhere is the thrill and mystery of
awakening life. That the animal is alive, and alive in a way you ought to be but
are not, is the last and most fascinating discovery you are likely to make in
nature’s kingdom. After years of intimate [36]observation, I can hardly meet a
wild bird or beast even now without renewed wonder at his aliveness, his
instant response to every delicate impression, as if each moment brought a
new message from earth or heaven and he must not miss it or the consequent
enjoyment of his own sensations. The very sleep of an animal, when he seems
ever on the thin edge of waking, when he is still so in touch with his changing

(Download PDF) Special Functions of Fractional Calculus Applications To Diffusion and Random Search Processes Trifce Sandev Full Chapter PDF

Uploaded by

Copyright:

Available Formats

You might also like

(Download PDF) Special Functions of Fractional Calculus Applications To Diffusion and Random Search Processes Trifce Sandev Full Chapter PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Download PDF) Special Functions of Fractional Calculus Applications To Diffusion and Random Search Processes Trifce Sandev Full Chapter PDF

Uploaded by

Copyright:

Available Formats

Special Functions of Fractional

Calculus: Applications to Diffusion and

Fractional Calculus and Fractional Processes with

Schaum's Outline of Probability, Random Variables, and

Micro-nanoelectronics Devices : Modeling of Diffusion

Calculus: Early Transcendental Functions 4th Edition,

Spatiotemporal Random Fields, Second Edition: Theory

Stochastic processes and random matrices : lecture

Handbook of solid state diffusion volume 1 Diffusion

Functions Modeling Change: A Preparation for Calculus,

12743_9789811252945_TP.indd 1 22/8/22 3:38 PM

This page intentionally left blank

B1948_1-Aoki.indd 6 9/22/2014 4:24:57 PM

12743_9789811252945_TP.indd 2 22/8/22 3:38 PM

Library of Congress Control Number: 2022945045

British Library Cataloguing-in-Publication Data

SPECIAL FUNCTIONS OF FRACTIONAL CALCULUS

ISBN 978-981-125-294-5 (hardcover)

For any available supplementary material, please visit

Desk Editor: Nur Syarfeena Binte Mohd Fauzi

Syarfeena - 12743 - Special Functions of Fractional Calculus.indd 1 25/8/2022 12:18:25 pm

To the memory of my father Miron-Meer Iomin,

This page intentionally left blank

B1948_1-Aoki.indd 6 9/22/2014 4:24:57 PM

Мы все учились понемногу,

Mathematics is the art of giving things misleading names.

Investigation of stochastic processes in complex media has been at-

viii Special Functions of Fractional Calculus

Skopje, Macedonia Trifce Sandev,

It is our great pleasure to thank our colleagues for collaboration, nu-

x Special Functions of Fractional Calculus

programme for experienced researchers at the Institute of Physics & As-

Acronyms and Symbols xv

2. Fox H-function and related functions 45

xii Special Functions of Fractional Calculus

2.2 Fox H-function . . . . . . . . . . . . . . . . . . . . . . . . 49

3. Elements of random walk theory 65

5. Heterogeneous diffusion processes 119

6. Diffusion processes with stochastic resetting 141

7. Random search 171

8. Diffusion on fractal tartan 185

xiv Special Functions of Fractional Calculus

8.2 Fractal mesh . . . . . . . . . . . . . . . . . . . . . . . . . 194

9. Finite-velocity diffusion 211

10. Appendices 229

Appendix A Functional calculus 231

Appendix B Stochastic differential equations 237

Appendix C Large deviation principle 243

Appendix D Fractals and fractal dimension 247

Appendix E Implementation of Wolfram Mathematica 251

Acronyms and Symbols

CTRW continuous time random walk

xvi Special Functions of Fractional Calculus

erf(z) error function

Acronyms and Symbols xvii

This page intentionally left blank

B1948_1-Aoki.indd 6 9/22/2014 4:24:57 PM

1.1 Integral transforms

In the limit → 0, integration over the contour C vanishes, and we obtain