Download as pdf or txt
Download as pdf or txt
You are on page 1of 53

Regional Analysis of Time Fractional

Diffusion Processes 1st Edition Fudong


Ge
Visit to download the full and correct content document:
https://textbookfull.com/product/regional-analysis-of-time-fractional-diffusion-processe
s-1st-edition-fudong-ge/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Fractional Diffusion Equations and Anomalous Diffusion


1st Edition Luiz Roberto Evangelista

https://textbookfull.com/product/fractional-diffusion-equations-
and-anomalous-diffusion-1st-edition-luiz-roberto-evangelista/

Fractional Calculus and Fractional Processes with


Applications to Financial Economics: Theory and
Application 1st Edition Fabozzi

https://textbookfull.com/product/fractional-calculus-and-
fractional-processes-with-applications-to-financial-economics-
theory-and-application-1st-edition-fabozzi/

Fractional Dynamic Calculus and Fractional Dynamic


Equations on Time Scales Svetlin G. Georgiev

https://textbookfull.com/product/fractional-dynamic-calculus-and-
fractional-dynamic-equations-on-time-scales-svetlin-g-georgiev/

Neutron diffusion : concepts and uncertainty analysis


for engineers and scientists 1st Edition Chakraverty

https://textbookfull.com/product/neutron-diffusion-concepts-and-
uncertainty-analysis-for-engineers-and-scientists-1st-edition-
chakraverty/
Time series analysis 1st Edition Palma

https://textbookfull.com/product/time-series-analysis-1st-
edition-palma/

GeoComputational Analysis and Modeling of Regional


Systems 1st Edition Jean-Claude Thill

https://textbookfull.com/product/geocomputational-analysis-and-
modeling-of-regional-systems-1st-edition-jean-claude-thill/

Functional and impulsive differential equations of


fractional order : qualitative analysis and
applications 1st Edition Stamov

https://textbookfull.com/product/functional-and-impulsive-
differential-equations-of-fractional-order-qualitative-analysis-
and-applications-1st-edition-stamov/

Design and analysis of time series experiments 1st


Edition Bartos

https://textbookfull.com/product/design-and-analysis-of-time-
series-experiments-1st-edition-bartos/

Regional Geology and Tectonics: Principles of Geologic


Analysis: Volume 1: Principles of Geologic Analysis 2nd
Edition Nicola Scarselli (Editor)

https://textbookfull.com/product/regional-geology-and-tectonics-
principles-of-geologic-analysis-volume-1-principles-of-geologic-
analysis-2nd-edition-nicola-scarselli-editor/
Fudong Ge · YangQuan Chen
Chunhai Kou

Regional Analysis
of Time-Fractional
Diffusion Processes
Regional Analysis of Time-Fractional Diffusion
Processes
Fudong Ge YangQuan Chen

Chunhai Kou

Regional Analysis
of Time-Fractional Diffusion
Processes

123
Fudong Ge Chunhai Kou
School of Computer Science Department of Applied Mathematics
China University of Geosciences Donghua University
Wuhan, Hubei Shanghai
China China

YangQuan Chen
Department of Mechanical Engineering
(MESA-Lab)
University of California, Merced
Merced, CA
USA

ISBN 978-3-319-72895-7 ISBN 978-3-319-72896-4 (eBook)


https://doi.org/10.1007/978-3-319-72896-4
Library of Congress Control Number: 2017961471

© Springer International Publishing AG 2018


This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by Springer Nature


The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
To Professor Abdelhaq El Jai
To our mentors and families
Preface

The twentieth century was rich in great scientific discoveries. One of the most
influential events is the introduction of diffusion process, which has been widely
used in physics (particle diffusion), chemistry, biology, sociology and economics
(diffusion of people, ideas, and price values), etc. Note that there exist two ways to
introduce the notion of diffusion process: (a) A phenomenological approach
according to Fick’s laws of diffusion; (b) A physical one by considering the random
walk of diffusing particles based on the theory of Brownian motion developed by
Albert Einstein in 1905. However, recently more and more experimental data and
field measurements have shown the dissatisfactions of Fick’s laws and the
assumptions introduced by Einstein due to the intrinsic heterogeneity media in these
diffusion processes. As a result, the conventional diffusion systems become unfit to
characterize these processes. Comparatively, these processes in heterogeneity media
are called to be anomalous diffusion processes, which recently were observed in a
wide variety of complex systems such as amorphous semiconductors, plasmas,
glassy materials, biological cells, epidemic spreading, etc. Furthermore, the data
measured from the anomalous processes satisfies a power-law function. Note that
Mittag-Leffler function also satisfies a power-law decay property and can be used to
express the solution of fractional differential equation. Based on the theory of
continuous-time random walks (CTRWs), it is proven that fractional diffusion
systems can be used to well characterize anomalous diffusion processes and offer
better performance not achievable before using conventional diffusion systems,
which raise many potential research opportunities at the same time.
As cited by El Jai and Pritchard in 1988, instead of analyzing a system by purely
theoretical viewpoint, using the notions of sensors and actuators to investigate the
structures and properties of systems could allow us to understand the systems better
and consequently, enable us to steer the real-world systems in a better way. This
situation happens in many real dynamic systems. Typical examples include the
optimal control of pest spreading, the swarm of robots moving through dense forest
or the flow through porous media microscopic process.
Besides, in the case of diffusion system, it is well known that in general, not all
the states can be reached/observed in the whole domain of interest. Here we

vii
viii Preface

introduce the notion of regional analysis, where the considered system is studied in
some subsets of the whole domain. It is worth mentioning that focusing on regional
analysis would allow for a reduction in the number of physical actuators/sensors,
offer the potential to reduce computational requirements in some cases, and also be
possible to discuss those systems which are not controllable/observable on the
whole domain.
What is the present book about? The aim of this book is to provide such an
overview of the regional control analysis of time fractional diffusion systems. The
models in this work are more or less academic and the results obtained are specific
to them. To the best of our knowledge, no results are available on these topics. We
hope that the results here could provide some insights into the study of fractional
diffusion systems.
This book consists of eight chapters.
In Chap. 1, briefly introduction of the CPSs, the DPSs and their new challenges
are given. Then, we state the concept of CTRW, its approach to anomalous dif-
fusion process and the advantages of using fractional equations to characterize the
CTRW’s guide to anomalous diffusion. An overview of the regional analysis is also
discussed.
Chapter 2 is concerned with some preliminary results to be used throughout this
book. To this end, we first study the gamma function, Mittag-Leffler functions, and
their properties. Then, we introduce the basic theory of fractional calculus, C0
semigroup and the Hilbert uniqueness methods, etc.
Chapters 3 and 4 are the main results of this monograph. In Chap. 3, the regional
controllability, regional gradient controllability, and regional boundary controlla-
bility of Riemann–Liouville-type and Caputo-type time fractional diffusion systems
on a time interval [0, b] are investigated. This is to find some control inputs, which
steer the considered system into a prescribed profile defined on a given subregion
of the system domain at time b. Some questions concerning with the number,
configurations of actuators and the minimum energy control approach are discussed
as well. In Chap. 4, it is realized that there is a need for further investigations on the
duality relationships between controllability and observability of the time fractional
diffusion systems. We discuss the regional observability, regional gradient
observability, and regional boundary observability for the Riemann–Liouville-type
and Caputo-type time fractional diffusion systems in a given subregion of the whole
domain. Besides, we answer the questions concerning the number, the configura-
tions of sensors, and the reconstruction approach of the initial vectors.
Most of the work in Chap. 5 is concerned with the problem of regional detection
of unknown sources for the Riemann–Liouville/Caputo-type time fractional diffu-
sion systems. More precisely, we introduce the notions of sources, detection, and
regional strategic/spy sensors. Then, the relationships between regional spy sensors
and regional strategic sensors to the Riemann–Liouville/Caputo-type time fractional
diffusion system are explored. Moreover, the problem on how to reconstruct a
regionally detectable source from the output equation is presented by using an
extensive HUMs approach.
Preface ix

In Chap. 6, we shall explore the spreadability of sub-diffusion processes, in


which the subdomains of the states to the system studied obeying a spatial property
are nondecreasing. These investigations pose rich in value to study those envi-
ronmental problems.
What’s more, Chap. 7 is concerned with the regional stability/stabilizability and
regional boundary stability/stabilizability to the Caputo/Riemann–Liouville-type
time fractional diffusion system. Some sufficient conditions to guarantee our main
results are shown and proven.
Finally, we present a survey of contributions of this monograph and given
several potential research directions, which are useful for engineering applications
and need for further investigations.
In summary, in this book, we introduce the regional idea and new methods to
tackle the CPSs, which are fantastic as well as challenging. During our research
processing, we frequently come across problems that are out of our expertise. We
are lucky that we can always gain valuable knowledge from our colleagues and
friends. We are grateful for all the people who had ever helped us and supported our
research. Some of them shared their valuable knowledge and experiences with us.
Many of them even spent their valuable time to review our work and provided their
faithful and insightful feedbacks.
Last but not least, we wish to thank Oliver Jackson, Editor, Springer UK, for his
visionary management that made this book possible. We also thank Nisha S
Keeran, Project Coordinator of Books Production with Springer Nature, Scientific
Publishing Services (P) Ltd. for her great patience and help in the copy editing
phase.

Wuhan, Hubei, China Fudong Ge


Merced, CA, USA YangQuan Chen
Shanghai, China Chunhai Kou
May 2017
Acknowledgements

Following Prof. Abdelhaq El Jai’s as well as Prof. El Hassan Zerrik’s pioneering


work on regional analysis of distributed parameter systems, this monograph iden-
tifies an emerging research topic by considering time fractional order diffusion
processes. With this Springer monograph, our intent is to provide a comprehensive
literature review and a summary of our research efforts during the past few years.
Naturally, this monograph contains materials from papers and articles that were
previously published. We are thankful and would like to acknowledge the copyright
permissions from the publishers who have released our work on that topic, as
specified in each chapter in question.

xi
Contents

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Cyber-Physical Systems and Distributed Parameter Systems . . . . . 1
1.1.1 Cyber-Physical Systems . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.2 Distributed Parameter Systems . . . . . . . . . . . . . . . . . . . . . 3
1.2 New Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Continuous Time Random Walk and Fractional Dynamics
Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3.1 Continuous Time Random Walk . . . . . . . . . . . . . . . . . . . . 7
1.3.2 Fractional Dynamics Approach . . . . . . . . . . . . . . . . . . . . . 8
1.4 Regional Analysis via Actuators and Sensors . . . . . . . . . . . . . . . . 9
1.4.1 Actuators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4.2 Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.4.3 Regional Analysis and the Emerging Research
Opportunities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 12
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 13
2 Preliminary Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...... 17
2.1 Special Functions and Their Properties . . . . . . . . . . . . . . ...... 17
2.1.1 The Gamma Function and Its Related Special
Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.1.2 The Mittag-Leffler Functions and Their Properties . . . . . . . 18
2.2 Fractional Calculus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2.1 Backgrounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2.2 Riemann–Liouville Fractional Integrals . . . . . . . . . . . . . . . 26
2.2.3 Riemann–Liouville Fractional Derivatives . . . . . . . . . . . . . 28
2.2.4 Caputo Fractional Derivatives . . . . . . . . . . . . . . . . . . . . . . 33
2.3 C0 Semigroups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.4 Hilbert Uniqueness Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

xiii
xiv Contents

3 Regional Controllability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 45


3.1 Regional Controllability . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 46
3.1.1 Riemann–Liouville-Type Time Fractional Diffusion
Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 46
3.1.2 Caputo-Type Time Fractional Diffusion Systems . . . ..... 61
3.2 Regional Gradient Controllability . . . . . . . . . . . . . . . . . . . ..... 79
3.2.1 Riemann–Liouville-Type Time Fractional Diffusion
Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 80
3.2.2 Caputo-Type Time Fractional Diffusion Systems . . . ..... 97
3.3 Regional Boundary Controllability . . . . . . . . . . . . . . . . . . . . . . . . 107
3.3.1 Riemann–Liouville-Type Time Fractional Diffusion
Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
3.3.2 Caputo-Type Time Fractional Diffusion Systems . . . . . . . . 115
3.4 Notes and Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
4 Regional Observability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
4.1 Regional Observability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
4.1.1 Riemann–Liouville-Type Time Fractional Diffusion
Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
4.1.2 Caputo-Type Time Fractional Diffusion Systems . . . . . . . . 131
4.2 Regional Gradient Observability . . . . . . . . . . . . . . . . . . . . . . . . . 138
4.2.1 Riemann–Liouville-Type Time Fractional Diffusion
Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
4.2.2 Caputo-Type Time Fractional Diffusion Systems . . . . . . . . 153
4.3 Regional Boundary Observability . . . . . . . . . . . . . . . . . . . . . . . . 167
4.3.1 Riemann–Liouville-Type Time Fractional Diffusion
Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
4.3.2 Caputo-Type Time Fractional Diffusion Systems . . . . . . . . 173
4.4 Notes and Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
5 Regional Detection of Unknown Sources . . . . . . . . . . . . . . . . . . . . . . 181
5.1 Preliminary Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
5.1.1 Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
5.1.2 Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
5.2 Riemann–Liouville-Type Time Fractional Diffusion Systems . . . . . 183
5.2.1 Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
5.2.2 Regional Strategic Sensors and Regional Spy Sensors . . . . 184
5.2.3 Reconstruction of a Regionally Detectable Source . . . . . . . 187
5.2.4 Several Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
5.3 Caputo-Type Time Fractional Diffusion Systems . . . . . . . . . . . . . 192
5.3.1 Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
5.3.2 The Relationships Between Strategic Sensors
and Spy Sensors in x . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Contents xv

5.3.3 An Approach to Regionally Detect the Source . . . . . . . . . 194


5.3.4 An Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
5.4 Notes and Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
6 Spreadability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
6.1 The Basic Knowledge of Spreadability . . . . . . . . . . . . . . . . . . . . 199
6.2 Riemann–Liouville-Type Time Fractional Diffusion Systems . . . . . 200
6.2.1 Spreading Control Problem . . . . . . . . . . . . . . . . . . . . . . . 201
6.2.2 Adaptive Spreading Control Problems . . . . . . . . . . . . . . . . 202
6.3 Caputo-Type Time Fractional Diffusion Systems . . . . . . . . . . . . . 207
6.3.1 Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
6.3.2 An Approach to Solve the Spreading Control Problem . . . . 208
6.4 Notes and Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
7 Regional Stability and Regional Stabilizability . . . . . . . . . . . . . . . . . 215
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
7.2 Regional Stability and Regional Stabilizability . . . . . . . . . . . . . . . 216
7.2.1 Caputo-Type Time Fractional Diffusion Systems . . . . . . . . 216
7.2.2 Riemann–Liouville-Type Time Fractional Diffusion
Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
7.3 Regional Boundary Stability and Regional Boundary
Stabilizability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
7.3.1 Caputo-Type Time Fractional Diffusion Systems . . . . . . . . 226
7.3.2 Riemann–Liouville-Type Time Fractional Diffusion
Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
7.4 Notes and Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
8 Conclusions and Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
8.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
8.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
8.2.1 Several Generalized Fractional Diffusion Systems . . . . . . . 234
8.2.2 Related Potential Research Opportunities . . . . . . . . . . . . . 238
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
Abbreviations

N Set of Natural Numbers


R Set of Real Numbers
Q Set of Rational Numbers
C Set of Complex Numbers
S? Orthogonal of S
♢* Adjoint Operator of Operator ♢
Re(z) Real Part of z
CPS Cyber-Physical System
CTRW Continuous-Time Random Walk
DPS Distributed Parameter System
FDS Fractional Diffusion System
HAB Harmful Algal Blooms
HUMs Hilbert Uniqueness Methods
Im(♢) Range of Operator ♢
Ker(♢) Kernel of Operator ♢
MIMO Multiple-Input Multiple-Output
MSD Mean Squared Displacement
NSF National Science Foundation
ODE Ordinary Differential Equation
PDE Partial Differential Equation
PDF Probability Density Function
SISO Single-Input Single-Output

xvii
List of Figures

Fig. 1.1 Measurement and control architecture of a CPS


(Reproduced with permission from Springer [62]) . . . . . . . . . . .. 2
Fig. 1.2 A comparative example between DPS and lumped parameter
system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 3
Fig. 1.3 Application scenario for wildfire control . . . . . . . . . . . . . . . . . . .. 5
Fig. 1.4 Application scenario for algal blooms monitoring
and control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Fig. 1.5 A simple CTRW model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Fig. 1.6 Illustration of the geometrical support to actuators . . . . . . . . . . . . 10
Fig. 1.7 Illustration of the spatial distribution to an actuator . . . . . . . . . . . 10
Fig. 1.8 Illustration of the geometrical support to sensors . . . . . . . . . . . . . 11
Fig. 2.1 Plots of the Mittag-Leffler function ea ðtÞ :¼ Ea ðta Þ
for a ¼ 0:25; 0:50; 0:75; 0:90; 1 respectively . . . . . . . . . . . . . . . .. 20
Fig. 2.2 Approximations of 1 þ Cð1aÞt 1
a ;
1
1 þ Cð1 þ aÞ1 ta
, and Ea ðta Þ
for a ¼ 0:35; 0:55; 0:75; 0:95. . . . . . . . . . . . . . . . . . . . . . . . . . . .. 21
Fig. 2.3 The statistics of academic publications on fractional calculus
[by Apr 2014] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 26
Fig. 3.1 Final reached state and target function on x  X at
time t ¼ 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 60
Fig. 3.2 Control input function, which is calculated by the formula
(3.1.50) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 60
Fig. 4.1 Regulation of the concentration flux of the substratum
at the upper bottom of the reactor . . . . . . . . . . . . . . . . . . . . . . . . . 138
Fig. 5.1 The difference between detection problem and observation
problem. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
Fig. 6.1 Spreads in S and in S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203

xix
Chapter 1
Introduction

In this chapter, we first introduce the concepts of CPSs, DPSs, and their new
challenges. Then, based on the theory of CTRW and its approach to anomalous
diffusion process, the advantages of using fractional order models to characterize the
CTRW’s guide to anomalous diffusion are presented.

1.1 Cyber-Physical Systems and Distributed Parameter


Systems

1.1.1 Cyber-Physical Systems

Cyber-physical systems (CPSs) are man-made complex systems coupled with nat-
ural processes that emerge as an integrative research field toward a new generation
of engineered systems. With integrated computational and physical processes, CPSs
can interact with humans through many new modalities [30]. Today, a precursor
generation of CPSs is found in areas as diverse as aerospace, automotive, chemical
processes, civil infrastructure, energy, health care, manufacturing, transportation,
entertainment, and consumer appliances [24]. However, to the best of our knowl-
edge, the term CPS has been given a lot of definitions and most of them depend
on the research fields of the people giving them. From a view of dynamic control
system, CPS can be defined as follows: “Computational thinking and integration of

Chapter 1 includes material ©2016 IEEE. Reprinted, with permission, from Fudong Ge,
YangQuan Chen, Chunhai Kou. “Cyber-physical systems as general distributed parameter sys-
tems: three types of fractional order models and emerging research opportunities,” IEEE/CAA
Journal of Automatica Sinica, Volume 2, Issue 4, Pages: 353–357. https://doi.org/10.1109/JAS.
2015.7296529.

© Springer International Publishing AG 2018 1


F. Ge et al., Regional Analysis of Time-Fractional Diffusion Processes,
https://doi.org/10.1007/978-3-319-72896-4_1
2 1 Introduction

What are
How the others
much ? saying ?
Should
I go ?
Actuator Sensor
Sensor
Physical World
Actuator
Sensor Actuator
Who is
available ?
What is
the best Where
spot ? to go?

Fig. 1.1 Measurement and control architecture of a CPS (Reproduced with permission from
Springer [62])

computation around the physical dynamic systems form CPSs where sensing, deci-
sion, actuation, computation, networking and physical processes are mixed.” This
definition can be effectively illustrated by Fig. 1.1 [62]. For more different definitions
of CPS, we refer the readers to [52, 59] defined by the National Science Foundation
(NSF), to [34] for the “networked” CPS definition, or to [60] for “wireless” CPS
definition, etc.
What’s more, since the objective of CPS is to develop new science and engineer-
ing methods, in which cyber and physical designs are compatible, synergistic, and
integrated at all scales [57, 62], advanced CPS will enable capability, adaptability,
scalability, resiliency, safety, security, and usability that will far exceed the simple
embedded systems. Precisely speaking, once the system knows the value of the phys-
ical variable either directly measured by sensors or indirectly calculated based on
the data from sensors, we can compare it with its target value. This comparison is
done by computing the difference, which is also called control deviation between
set point and current value. The control deviation is forwarded to a controller, which
then adjusts its outputs according to its inherent control strategy to bring the system
more closely to the target set point. The output controls an actuator which influences
the physical behavior of the system under control. Moreover, note that the applica-
tions of CPS are rich in our real world. Typical examples include medical devices
systems, automotive and air traffic control, advanced automotive systems, process
control, environmental monitoring, avionics, instrumentation, oil refineries, water
usage control, cooperative robotics, manufacturing control, smart greener buildings,
etc. We believe that CPS technology would totally change the way people interact
with engineered systems—just as the Internet has changed the way people interact
with information.
1.1 Cyber-Physical Systems and Distributed Parameter Systems 3

Fig. 1.2 A comparative example between DPS and lumped parameter system

However, the challenge is that during the control system design processing, CPSs
are common across multiple application domains, including the real-time control,
the user software applications along with opportunities for accelerated transition to
practice [41]. Besides, in many control systems, the user software application imple-
ments a man–machine interface for system status visualization and for management
of certain operational parameters, in which the closed-loop real-time control portion
must regulate the environment with minimum jitter at all given time. Then things get
more complicated [30]. Therefore, some model-based control strategies are needed
to steer the CPSs to a desired behavior.
From a view of mathematics, distributed parameter systems are used efficiently
to study CPSs due to the fact that CPSs are mostly dynamic evolutions arising not
only along the time axis but also along spatial axis.

1.1.2 Distributed Parameter Systems

Distributed parameter systems (DPSs) are dynamical systems in which all depen-
dent variables are functions of time and one or more spatial variables. In this case,
DPSs become infinite-dimensional dynamical systems usually governed by partial
differential equations, delay differential equations and some differential equations in
abstract spaces, etc. Conversely, if the dependent variables of the dynamical system
are a function of time alone, we call it the lumped parameter system (LPS), which
in general, is a finite-dimensional dynamical system modeled by ODEs.
Strictly speaking, distribution exists naturally in any physical systems and DPS is
the inevitable development and generalization of LPS. To study a real-world problem
governed by a LPS can be viewed as a approximation of that by a DPS. Therefore,
if we intend to accurately describe a practical application, the DPS is necessary. For
example, see Fig. 1.2. Obviously, system (a) should be governed by a DPS, consisting
of an infinitely thin string, supported at both ends. The dependent variable, the vertical
position of the string y(x, t) is indexed continuously in both space and time. While,
the system (b) consists of a series of “beads” connected by massless string segments,
constrained to move vertically and can be regarded as a reduced approximation.
Then, system (b) is thought of as a lumped parameter system governed by n−order
ODEs [5].
4 1 Introduction

However, due to the poor tools for the theory analysis of infinite-dimensional
dynamic systems, the development of DPS is very slow until the middle of last cen-
tury when the theory of functional analysis, differential dynamic system is initially
explored. By using the concept of transfer function, the heat conduction problems
governed by DPS are discussed in [51] and in [7], the optimal control of DPS is
studied by using the moments analysis methods. We refer the readers to [65] and the
references therein for an initial summarization of the basic theory of DPS including
its stability, controllability, observability, etc. In 1970s, the theories of functional
analysis, partial differential equation, and operator semigroup witnessed a huge
development, which provided many powerful tools to study DPS. Here, we refer
the readers to monographs [35, 36] for a pioneering work on the optimal control
of DPS and to [56] for several well-identified research directions in the study of
DPS including optimal control, measurement, model reduction, numerical methods,
etc. Furthermore, many breakthrough contributions are made in the fields of optimal
location of sensors/actuators [26, 28, 48], variable identification [27, 29], adaptive
control [6, 14], compensator design [54], stability theory analysis [64], fuzzy control
[67], etc. In a word, the field of mathematics encompassing DPS has experienced a
great development during the period of last century. To the best of our knowledge,
there exist two main methods for the study of DPSs:
• Investigate DPSs by studying the properties of partial differential equations, such
as diffusion equation, reaction–diffusion equation, etc.;
• Study DPSs by utilizing the theory of infinite-dimensional dynamic system, func-
tional analysis, operator semigroup, etc.
In the new twenty-first century, there is a growing emphasis on the applications
of DPS, which is gradually applied in the fields of aerospace, biology, chemistry,
civil engineering, electronics, mechanical engineering, etc. The following is a short
collection of them.
• Fluid dynamics [3];
• Heat dynamics [37];
• Soil dynamics [50];
• Wildlife damage management dynamics [31];
• Signal transmission lines dynamics [16];
• Dynamic environmental monitoring [13, 46];
• Electromagnetic suspension dynamics [55];
• etc. · · ·
For a more complete review of DPS applications, we refer the readers to [33, 47, 53]
and the references therein. Here, we describe two of them.
(1) Wildfire Control Scenario
Detecting and controlling wildfires in its early stage are very significant to protect
forests especially during the dry season. See Fig. 1.3, a group of UAVs is sent to detect
the boundary of the fire, meanwhile, firemen are dispatched in the area surrounding
the fire, waiting for instructions. DPS models combined with information such as
1.1 Cyber-Physical Systems and Distributed Parameter Systems 5

Wind Humidity

Fig. 1.3 Application scenario for wildfire control

Hurricane

Density
Temperature

Fig. 1.4 Application scenario for algal blooms monitoring and control

the speed direction of wind, the humidity of forest and the source location of fire can
be used to decide when and how to send fire trucks. For more scenarios on wildfire
control scenario, please refer to [31, 62].
(2) Algal Blooms Monitoring and Control Using Mobile Actuators/Sensors
Harmful Algal Blooms (HAB) are a menace to water wildlife as the release of toxins
into habitats that can generate large population death count. However, the lack of
systematic approach to detect, forecast, and control HAB means that most scientists
study their aftermaths rather than their preventions. So far, scientists used poorly
calibrated tools for their problems. See Fig. 1.4 [62], by using information from UAVs
and the DPS models, the actuators could be sent, so that the release of mitigating
agents would have the most impact on either the population of algae or harmful
6 1 Introduction

chemical. For more information on how to monitor and control the algal blooms, we
refer the readers to [62] and the references therein.

1.2 New Challenges

As we all know, Fick’s law is the standard law describing diffusion processes in
homogeneous media. However, recently more and more experimental data and field
measurements have shown strong dissatisfaction of the Fick’s law due to the intrinsic
heterogeneity media in these diffusion processes. As a result, the traditional diffusion
systems become unfit to characterize these processes. Comparatively, these processes
in heterogeneity media are called to be anomalous diffusion processes, which are
ubiquitous in the natural sciences and show properties of memory and hereditary
effects, path dependence, and global connection. We summarize the characteristics
of anomalous diffusion processes as follows:
1. The anomalous diffusion process typically contains a large diversity of elementary
units and strong interactions existing among these basic units;
2. The anomalous evolution is non-predictable as time evolves. In general, the time
evolution of such process deviates from the corresponding standard laws;
3. The anomalous diffusion process is a nonlocal process and does not satisfy the
Fourier’s heat conduction law and the Fick’s diffusion law.
In fact, many complex real-world dynamical systems often contain anomalous dif-
fusion properties, for example, the fluctuation of price in financial market, the move-
ment of pollutants in the natural environment, the movement of plasma under high
temperature and high pressure, or the diffusion, dissipation, and penetration of soft
materials. From a view of physical modeling, the classical integer order derivative is a
local operator, which is not suitable to characterize the nonlocal processes. However,
the fractional derivative operator is actually a integro-differential convolution oper-
ator and exhibits (weakly singular) kernels of the power-law type. This means that
fractional derivative operator can be used to positively reflect the dynamic processes
with properties of memory and hereditary effects, path dependence or global con-
nection, etc. [39, 61, 66]. Hence, the fractional diffusion systems provide a natural
description of the nonlocal transportation and then show better performance.
Recently, anomalous diffusion modelings have attracted increasing interest and
a lot of contributions have been given to them [4, 10, 38, 43] especially after the
introduction of CTRW [45] due to their closed connections with time fractional dif-
fusion equations [22]. This is because that data measured from anomalous diffusion
processes satisfy a power-law function and so does the solution of fractional differ-
ential equations. For further information on this, we refer the readers to Figs. 2.1 and
2.2 below for the power-law property of Mittag-Leffler function and the fact that
Mittag-Leffler function can be used to express the solution of fractional diffusion
systems [18, 25]. Then based on these, together with the existing research results on
1.2 New Challenges 7

distributed parameter systems, the theory of infinite-dimensional dynamic system,


and the theory of fractional calculus and fractional differential equations, we shall
study the regional controllability, regional observability and spreadability, etc., of
the time fractional diffusion systems. To the best of our knowledge, no results are
available on these topics. We hope that the results here could provide some insights
into the control theory of this field and be used in real-life applications.

1.3 Continuous Time Random Walk and Fractional


Dynamics Approach

In this section, we first introduce the concept of CTRW and its approach to anomalous
diffusion process. Then, we present the advantages of using fractional equations to
characterize the CTRW’s guide to anomalous diffusion.

1.3.1 Continuous Time Random Walk

Since the mean squared displacement (MSD) of anomalous diffusion is the power law
of a fractional exponent compared with a linear function in the Gaussian diffusion
process [39, 61], we see that the Brownian motion will fail to characterize anomalous
diffusion process. Here, we shall introduce a new method to approach them—the
CTRW approach.
Regarded as a random walk subordinated to a simple renewal process, CTRWs,
governed by a waiting time probability density function (PDF) and a jump length
PDF, have been proven to be a generalization tool of physical diffusion process to
effectively describe the anomalous diffusive transports in our real life [4, 10, 43,
44]. See Fig. 1.5 for a simple example, where the walker jumps at each time step to a

Fig. 1.5 A simple CTRW


model t
t3

t2

t1

x
x2 x0 x1
8 1 Introduction

randomly selected direction. More precisely, when the particles are assumed to jump
at fixed time intervals with incorporating waiting times, the particles will experience
the sub-diffusion processes, such as the flow through porous media microscopic
process [63], or the swarm of robots moving through dense forest [58], etc. In this
case, the MSD is a power law of fractional exponent smaller than that of a Gaussian
diffusion process. Besides, when the particles are supposed to jump following from
a general, non-Gaussian jump distribution function, the particles then undergo the
super-diffusion process [10, 21]. Now, the MSD is a power-law function of fractional
exponent bigger than that of a Gaussian diffusion process.

1.3.2 Fractional Dynamics Approach

By [22], there exists a closed connection between CTRWs and the diffusion equations
with fractional time derivatives. It is confirmed in [23, 39, 43] that the diffusion
equations with fractional time derivatives could be used to well describe the sub-
diffusion processes, which surely raise many potential research opportunities at the
same time.
Now, we state the answers of the following two questions:
(1) What can fractional equations do?
For the fractional calculus, as we all know, it has shown great potential in science
and engineering applications. Some phenomena such as self-similarity, nonstation-
ary, non-Gaussian process, and short or long memory process are all closely related
to fractional calculus [2, 25, 32, 49]. Then fractional diffusion systems could be used
to provide a natural description of nonlocal transport so as to well characterize the
anomalous processes [4, 43].
(2) What can fractional equations do better to approach CTRW’s guide to
anomalous diffusion?
(a) The essential advantages: Compared with Gaussian diffusion process, frac-
tional diffusion process is nonuniversal in that involves a parameter α, which is the
order of the fractional derivative. Then, fractional diffusion equations account for
the typical “anomalous” features which are observed in many real-life systems.
(b) The comparisons to other approaches: Compared to other approaches, the frac-
tional approach is in some sense equivalent to the generalized conventional equation
approach, which with time-dependent coefficients may lead to a different description
corresponding to conventional equations that are local in time. For example, in most
cases of Brownian transports, the deterministic Fokker–Planck equation is employed
for the description of stochastic dynamics in external fields. Similarly, we promote
to use the fractional Fokker–Planck equation for situations where anomalous diffu-
sion underlies the system. Then, the standard techniques for solving conventional
differential equations also maybe useful to the fractional cases.
(c) The mathematical advantages: The appearance of fractional equations is
very appealing due to their proximity to the analogous standard equations. Another
1.3 Continuous Time Random Walk and Fractional Dynamics Approach 9

advantage of the fractional model may lie in the straightforward way of including
external force terms and of calculating boundary value problems.
All in all, it is now widely believed that, using fractional calculus in modeling is
not just another way of presenting old stories, it could better capture the complex
dynamics of natural and man-made systems, and that could offer better performance
not achievable before using integer order controls [39, 61], which in fact raise impor-
tant theoretical challenges and open new research opportunities.

In this literature, we shall use the time fractional diffusion system where the
conventional first-order time derivative is extended to a fractional derivative of order
α ∈ (0, 1] to well characterize the sub-diffusion process. For more information on
the advantages of our approaches, we refer the readers to papers [43, 44]. By relating
our and others’ work in that field and putting it in some more general context, we
believe that this present monograph may be the basis for some active researches on
complex dynamics by using a tool as old as classical calculus.

1.4 Regional Analysis via Actuators and Sensors

The sensors and actuators are the intermediaries between a real system and its envi-
ronment. The actuators permit the excitation of the system and can be of many
different forms. The sensors play a passive role and enable one to gather informa-
tion on the system and its evolution; these will be studied in a later chapters. In this
section, we shall present the definitions on configurations of sensors and actuators
as well as the class of the systems to be considered in this monograph.
As stated in [15, 17], instead of analyzing a system by purely theoretical viewpoint
(see [12] for example), using the notions of sensors and actuators to investigate the
structures and properties of systems could allow us to understand the system better
and consequently enable us to steer the real-world system in a better way. This
situation happens in many real dynamic systems, for example, the optimal control
of pest spreading [8], the crowd-pedestrian egress and evacuation [9], etc.

1.4.1 Actuators

Let Ω ⊆ Rn be an open and bounded subset with a smooth boundary Γ = ∂Ω.


Definition 1.4.1 (i) An actuator is a couple (D, g) where D = supp(g) ⊂ Ω repre-
sents the geometrical support of the actuator, g is its spatial distribution.
(ii) An actuator (D, g) is said to be
• a zone actuator if D is a nonempty subregion of Ω;
• a pointwise actuator if D is reduced to a point σ ∈ Ω. In this case, we have g = δσ
where δσ is the Dirac function concentrated at σ . The actuator is denoted as (σ, δσ );
10 1 Introduction

Fig. 1.6 Illustration of the


geometrical support to
actuators

Fig. 1.7 Illustration of the


spatial distribution to an
actuator

• a boundary actuator if its support D ⊂ Γ . In this case, it could also be divided into
zone actuator and pointwise actuator.

Moreover, we refer the readers to Figs. 1.6 and 1.7 respectively [15], for an illustration
of the geometrical support and the spatial distribution of actuators.
In Definition 1.4.1, we assume that g ∈ L 2 (D). In the case of p actuators
(Di , gi )1≤i≤ p , we have U = R p and

B : Rp → L 2 (Ω)
p
(1.4.1)
u(t) → Bu(t) = gi u i (t),
i=1

where t ∈ [0, b], u = (u 1 , u 2 , ..., u p )T ∈ 


L 2 (0, b; R p ) and gi ∈ L 2 (Di ) with Di =
supp(gi ) ⊂ Ω for i = 1, 2, . . . , p and Di D j = ∅ for i = j, and we have

B  y = (< g1 , y >, . . . , < g p , y >)T for z ∈ L 2 (Ω), (1.4.2)

where M T is the transpose matrix of M and ·, · is the inner product in L 2 (Ω).


1.4 Regional Analysis via Actuators and Sensors 11

Fig. 1.8 Illustration of the


geometrical support to
sensors

Furthermore, we state that if D is not dependent on time, the actuator (D, g) is


said to be fixed or stationary. Otherwise, it is a moving or mobile actuator denoted
by (Dt , gt ) where Dt and gt are, respectively, the geometrical support and the spatial
distribution of the actuator at time t.

1.4.2 Sensors

We introduce hereafter the notion of sensors for DPS.


Definition 1.4.2 A sensor is a couple (D, h) where D = supp(h) ⊂ Ω is the support
of the sensor, and h is its spatial distribution.
We refer the readers to Fig. 1.8 for an illustration of the geometrical support of sensors
[15].
Generally, we assume that h ∈ L 2 (D). Then we can similarly define zone or
pointwise, internal or boundary, fixed or moving sensors. If the output of the system is
 (Di , h i )1≤i≤q with h i ∈ L (Di ), Di = supp(h i ) ⊂
2
given by means of q zone sensors
Ω for i = 1, 2, . . . , q and Di D j = ∅ if i = j, then in the zone case, the DPS’s
output operator C is defined by

C : L 2 (Ω) → Rp
(1.4.3)
y → C y = (< h 1 , y >, . . . , < h q , y >)T

and the output is given by


⎡ ⎤
< h 1 , y > L 2 (D1 )
⎢ < h 2 , y > L 2 (D2 ) ⎥
⎢ ⎥
z(t) = ⎢ .. ⎥. (1.4.4)
⎣ . ⎦
< h q , y > L 2 (Dq )
12 1 Introduction

A sensor (D, h) is said to be a zone sensor if D is a nonempty subregion of Ω.


The sensor (D, h) is pointwise if D is reduced to a point c ∈ Ω, and in this case
h = δc is the Dirac function concentrated at c. The sensor is denoted as (c, δc ). In
the case of zone or pointwise sensors, if D ⊂ Γ = ∂Ω, the sensor (D, h) is said to
be a boundary sensor. If D is not dependent on time, the sensor (D, h) is said to be
fixed or stationary; otherwise it is said to be moving (or a scanning) sensor and is
denoted as (Dt , h t ). In the case of q pointwise fixed sensors located in (ci )1≤i≤q , the
output function is a q-vector given by
⎡ ⎤
y(t, c1 )
⎢ y(t, c2 ) ⎥
⎢ ⎥
z(t) = ⎢ .. ⎥, (1.4.5)
⎣ . ⎦
y(t, cq )

where ci is the location of the sensor i and y(t, ci ) is the state of the system in ci at
time t.
Moreover, while the actuators allow excitations on the system, the sensors have
a passive role of giving information (measurements) on the system and its state
evolution. The actuators and sensors can be of different natures: zone or pointwise,
internal or boundary, stationary or moving, communicating or noncommunicating,
collocated or non-collocated.
Besides, from engineering point of view, using the notions of actuators and sensors
to tackle DPSs would allow us to describe more concretely the exchanges between
system and its environment. The study can then be considered, not only with respect
to the operators in system but also with respect to the spatial distribution, location,
and number of the actuators and sensors, which in fact raise many new research
opportunities.

1.4.3 Regional Analysis and the Emerging Research


Opportunities

1.4.3.1 Regional Analysis

Note that the controllability problem of a fractional order sub-diffusion system can
be reformulated as a problem of infinite-dimensional control system. In the case of
diffusion systems, however, in general, not all the states can be reached or observed in
the whole domain of interesting [1, 15, 18–20]. Then the regional analysis idea should
be introduced. Not only could regional analysis be more efficient to guide the practical
applications, focusing on regional analysis would also allow for a reduction in the
number of physical actuators/sensors and offer the potential to reduce computational
1.4 Regional Analysis via Actuators and Sensors 13

requirements in some cases. Besides, it could be possible to discuss the systems which
are not controllable/observable on the whole domain with the concept of regional
analysis.

1.4.3.2 The Emerging Research Opportunities

As cited above, fractional order DPSs have now been found wide applications in
describing many physical phenomena. At the same time, to our best knowledge,
many problems are still open calling for research cooperation of multidisciplines
such as mathematical modeling, engineering applications, and information sciences.
Potential topics such as modeling the anomalous diffusion process and its control
using fractional order DPSs, in particular for actuator and sensor networks, do arise
in numerous robotic and eventually automotive applications [40, 42]. These include
the characterizations of the actuators/sensors when the systems under consideration
are regional controllability/observability, distributed control, and source detection as
well as issues such as robustness and optimality. On the one hand, fractional calculus
will enable to introduce novel modeling and control techniques for the anomalous
diffusion processes. The consideration is motivated by observing that many con-
sensus algorithms for distributed systems are designed in the continuous form. This
kind of control design is independent of the actual communication topology that
is imposed by means of discretization. As a result, fractional order DPS provides
a new perspective to address the arising control problems for anomalous diffusion
processes.
Furthermore, in the case of diffusion systems, as mentioned above, due to the
spatial–temporal sampling and discrete nature of decision and control, the notions
of regional analysis should be introduced, i.e., we can consider the regional stability,
regional controllability, regional observability, etc., of the system under considera-
tion. In addition, by [11], from an application point of view, some plain questions
such as “How many actuators are used and how to configure them so that the system
studied can behave (hopefully) according to our wishes?”, “If there are many controls
that steer the system to a desired state in a given region ω, is it possible to find a way
to approach the minimum energy cost? If yes, how does it depend on ω and on the
locations of the controls?”, “If the system studied is regionally observable/gradient
observable/boundary observable, how to reconstruct the initial vector and it relation-
ship with locations of sensors?”, etc. might be asked, which in fact raise important
theoretical challenges and open new research opportunities.

References

1. Afifi L, Chafia A, Jai AE (2002) Regionally efficient and strategic actuators. Int J Syst Sci
33(1):1–12
14 1 Introduction

2. Agrawal OP (2002) Solution for a fractional diffusion-wave equation defined in a bounded


domain. Nonlinear Dyn 29(1–4):145–155
3. Batchelor GK (2000) An introduction to fluid dynamics. Cambridge University Press, Cam-
bridge
4. Benson DA, Wheatcraft SW, Meerschaert MM et al (2000) The fractional-order governing
equation of Lévy motion. Water Resour Res 36(6):1413–1423
5. Bilbao S, Smith III JO (2014) MUS420/EE367A lecture 7D: discrete-time lumped models.
In: Center for computer research in music and acoustics (CCRMA), Department of Music,
Stanford University. March 24, 2014. https://ccrma.stanford.edu/~jos/NumericalInt/
6. Bohm M, Demetriou MA, Reich S, Rosen IG (2006) Model reference adaptive control of
distributed parameter systems. SIAM J Control Optim 36(1):33–81
7. Butkovsky A (1961) Some approximate methods for solving problems of optimum control of
distributed parameter systems. Avtomat ITelemekh 22(12):1565–1575
8. Cao J, Chen Y, Li C (2015) Multi-UAV-based optimal crop-dusting of anomalously diffusing
infestation of crops. In: American Control Conference Palmer House Hilton July 1–3, 2015.
Chicago, IL, USA. See also: arXiv:1411.2880
9. Cao K, Chen Y, Stuart D, Yue D (2015) Cyber-physical modeling and control of crowd of
pedestrians: a review and new framework. IEEE/CAA J Autom Sin 2(3):334–344
10. Cartea A, Negrete D (2007) Fluid limit of the continuous-time random walk with general Lévy
jump distribution functions. Phys Rev E 76(4):1–8
11. Chen Y, Moore KL, Song Z (2004) Diffusion boundary determination and zone control via
mobile actuator-sensor networks (MAS-net): challenges and opportunities. In: Defense and
security, international society for optics and photonics, pp 102–113
12. Curtain RF, Zwart H (2012) An introduction to infinite-dimensional linear systems theory,
vol 21. Springer Science & Business Media, Berlin
13. Dereszynski EW, Dietterich TG (2011) Spatiotemporal models for data-anomaly detection in
dynamic environmental monitoring campaigns. ACM Trans Sens Netw (TOSN) 8(1):3
14. Duncan TE, Maslowski B, Pasikduncan B (1994) Adaptive boundary and point control of linear
stochastic distributed parameter systems. SIAM J Control Optim 32(3):648–672
15. El Jai A, Pritchard AJ (1988) Sensors and controls in the analysis of distributed systems. Halsted
Press, Europe
16. Futagami T, Tzafestas SG, Sunahara Y (1989) Distributed parameter systems-modelling and
simulation. In: Proceedings of the IMACS-IFAC international symposium. Elsevier Science
Inc, Amsterdam
17. Ge F, Chen Y, Kou C (2017) Actuator characterisations to achieve approximate controllability
for a class of fractional sub-diffusion equations. Int J Control 90(6):1212–1220
18. Ge F, Chen Y, Kou C (2016) Regional boundary controllability of time fractional diffusion
processes. IMA J Math Control Inf, 1–18
19. Ge F, Chen Y, Kou C (2016) Regional gradient controllability of sub-diffusion processes. J
Math Anal Appl 440(2):865–884
20. Ge F, Chen Y, Kou C (2017) Regional controllability analysis of fractional diffusion equations
with Riemann-Liouville time fractional derivatives. Automatica 76:193–199
21. Gradenigo G, Sarracino A, Villamaina D, Vulpiani A (2013) Einstein relation in systems with
anomalous diffusion. Acta Phys Polo B 44(5):899–912
22. Hilfer R, Anton L (1995) Fractional master equations and fractal time random walks. Phys Rev
E 51(2):848–851
23. Hilfer R, Butzer P, Westphal U, Douglas J, Schneider W, Zaslavsky G, Nonnemacher T, Blumen
A, West B (2000) Applications of fractional calculus in physics, vol 128. World Scientific,
Singapore
24. Khaitan SK, McCalley JD (2015) Design techniques and applications of cyberphysical systems:
a survey. Syst J IEEE 9(2):350–365
25. Kilbas AA, Srivastava HM, Trujillo JJ (2006) Theory and applications of fractional differential
equations. Elsevier Science Limited, Amsterdam
References 15

26. Kim M, Erzberger H (1967) On the design of optimum distributed parameter system with
boundary control function. IEEE T Automat Contr 12(1):22–28
27. Kravaris C, Seinfeld JH (1985) Identification of parameters in distributed parameter systems
by regularization. SIAM J Control Optim 23(2):217–241
28. Kubrusly C, Malebranche H (1985) Sensors and controllers location in distributed systems: a
survey. Automatica 21(2):117–128
29. Kubrusly CS (1977) Distributed parameter system indentification: a survey. Int J Control
26(4):509–535
30. Lee E (2008) Cyber physical systems: design challenges. In: 11th IEEE international sympo-
sium on object oriented real-time distributed computing (ISORC), 2008. IEEE, New York, pp
363–369
31. Lenhart SM, Bhat MG (1992) Application of distributed parameter control model in wildlife
damage management. Math Models Methods Appl Sci 2(04):423–439
32. Li B, Wang J (2003) Anomalous heat conduction and anomalous diffusion in one-dimensional
systems. Phys Rev Lett 91(4):044,301
33. Li HX, Qi C (2010) Modeling of distributed parameter systems for applicationsła synthesized
review from time-space separation. J Process Control 20(8):891–901
34. Liberatore V (2007) Networked cyber-physical systems: an introduction. In: Proceedings of
Mobisensors 07
35. Lions JL (1971) Optimal control of systems governed by partial differential equations, vol 170.
Springer, Berlin
36. Lions JL, Ageneies E (1972) Non-Homogeneous boundary value problems and application.
Springer, New York
37. Madsen H, Holst J (1995) Estimation of continuous-time models for the heat dynamics of a
building. Energy Build 22(1):67–79
38. Mainardi F, Luchko Y, Pagnini G (2001) The fundamental solution of the space-time fractional
diffusion equation. Fract Calc Appl Anal 4(2):153–192
39. Mandelbrot BB (1983) The fractal geometry of nature, vol 173. Macmillan, London
40. Martinoli A, Mondada F, Correll N, Mermoud G, Egerstedt M, Hsieh MA, Parker LE, Støy
K (2012) Distributed autonomous robotic systems: the 10th international symposium, vol 83.
Springer, Berlin
41. Marwedel P (2010) Embedded system design: Embedded systems foundations of cyber-
physical systems. Springer Science & Business Media, Berlin
42. Mesbahi M, Egerstedt M (2010) Graph theoretic methods in multiagent networks. Princeton
University Press, Princeton
43. Metzler R, Klafter J (2000) The random walk’s guide to anomalous diffusion: a fractional
dynamics approach. Phys Rep 339(1):1–77
44. Metzler R, Klafter J (2004) The restaurant at the end of the random walk: recent developments
in the description of anomalous transport by fractional dynamics. J Phys A 37:161–208
45. Montroll EW, Weiss GH (1965) Random walks on lattices. II. J Math Phys 6(2):167–181
46. Nittel S (2009) A survey of geosensor networks: advances in dynamic environmental monitor-
ing. Sensors 9(7):5664–5678
47. Omatu S, Seinfeld JH (1989) Distributed parameter systems: theory and applications. Claren-
don Press, UK
48. Omatu S, Koide S, Soeda T (1978) Optimal sensor location problem for a linear distributed
parameter system. IEEE Trans Autom Control 23(4):665–673
49. Podlubny I (1999) Fractional differential equations, vol 198. Academic Press, Cambridge
50. Prakash S (1981) Soil dynamics. McGraw-Hill Companies, New York
51. Qian X (1954) Engineering cybernetics. McGraw-Hill, New York
52. Ramaprasad H (2008) Providing end-to-end guarantees in cyber-physical systems. Southern
Illinois University Carbondale, Technical report
53. Ray W (1978) Some recent applications of distributed parameter systems theoryła survey.
Automatica 14(3):281–287
Another random document with
no related content on Scribd:
[445] This is all we are told by Ibn al Athîr. But there is
elsewhere a not unlikely tradition that the unhappy maiden,
tearing herself from her captor’s embrace, leapt from the camel,
and found in death an escape from her humiliation. This
campaign furnishes plentiful material for many still wilder stories
in the romances of the pseudo-Wâckidy and later writers.
[446] According to some authorities, Othmân presented the
royal share of the booty as a free gift to Merwân, and they add
that this was one of the grounds of Othmân’s impeachment. But it
reads like a party calumny.
[447] Coming there in disguise, he was recognised by a
woman, who gave the alarm, and the natives rushed upon the
boat. Asked how she recognised the Saracen captain, this
woman said, ‘He came as a merchantman; but when I asked an
alms of him, he gave as a prince giveth; so I knew it was the
captain of the Saracens.’
The payment of jazia, or poll-tax, implied the corresponding
claim of protection. Zimmy signifies one who, so assessed,
becomes part and parcel of the Moslem empire, and as such
entitled to its guardianship. The Cypriots were not expected, from
their position, to take any active part on the Moslem side; but they
were bound to give their new masters warning of any hostile
expedition, and generally to facilitate their naval operations.
As the great crowd of prisoners were being shipped, one of
the Moslem warriors wept; for, said he, ‘those captives will lead
the hearts of their masters astray’—one of the few occasions on
which we see a faint perception of the evils of female slavery to
the conquerors themselves; for that I take to be the meaning.
[448] According to Theophanes, it was Constans II., grandson
of Heraclius, who perished thus for his crimes, but at a later date.
See Gibbon, ch. xlviii.
[449] Some authorities make the discontent to arise in
consequence of the failure of Abu Sarh to follow up the victory,
and give chase to the retiring enemy.
[450] See Life of Mahomet, p. 235.
[451] Abbasside tales are multiplied against the unfortunate
Welîd. He consorted with the poet Abu Zobeid, a converted
Christian of the Beni Taghlib, and was suspected of drinking wine
in company with him. Another complaint was, that a conjuring Jew
from Baghdad having been condemned in Ibn Masûd’s court for
witchcraft, Jondob, one of the factious leaders, killed him with his
own hand instead of waiting the regular course of execution; for
which unlawful act Welîd imprisoned Jondob, to the great
discontent of the people. Hostile tradition, by deep colouring, has
improved on these tales, representing Welîd as a brutal sot and
sacrilegist. E.g. by his command, the Jew performed works of
magic in the sacred precincts of the Great Mosque, assumed by
sorcery the form of various animals, cut off a man’s head, and
then putting it on, brought him to life again, &c. Jondob,
scandalised at his devilish tricks, proceeded to cut off the Jew’s
head, saying, ‘If thou canst do miracles, then bring thyself to life
again.’ Upon this Welîd imprisoned Jondob, and would have put
him to death had he not, by the connivance of the jailor, escaped.
These tales are given by Masûdi and later writers, whose
tendency to vilify Welîd by the most extravagant fiction, is
manifest.
Of the same complexion are the traditions which represent the
citizens of Medîna as in such bodily fear of Othmân that no one
dared to carry out the sentence of scourging against Welîd; so
that Aly, at last, stepped forward, and himself inflicted the stripes.
Others say that Aly ordered his son, Hasan, to do so; but he
refused, saying, ‘The lord of the hot is lord also of the cold’ (i.e.
the sweets and the bitter of office must go together), and that then
Aly compelled a grandson of Abu Tâlib (Mahomet’s uncle) to carry
out the sentence.
[452] His name was Abdallah, but to distinguish him from the
multitude of that name, he is always called Ibn Aámir.
[453] The youth, however, was not satisfied with this pair of
wives; for he left twenty sons, and as many daughters, behind
him. He was nephew of that Khâlid ibn Saîd who opened the
Syrian campaign so ingloriously.
[454] On the text of the Corân, and the history of this
recension, see the Excursus on the ‘Sources for the Biography of
Mahomet,’ in the Life of Mahomet. The manner in which the
Abbasside faction perverted the facts and turned the charge to
malignant purpose against the Omeyyad house, will be
understood from the section on the Corân in the Apology of Al
Kindy (Smith and Elder, 1882), pp. 25 et seq. The charge against
Al Hajjâj of having altered the text is equally groundless. See Ibid.
p. xi.
[455] The precise nature of the arrangement, as stated by Ibn
al Athîr, is not very clear, but its general character seems to have
been as given the text.
[456] Masûdy, an unprejudiced witness, dwells on this as one
of the causes of demoralisation and disloyalty now setting in so
rapidly, and he gives some remarkable instances. Zobeir had
1,000 slaves, male and female, and 1,000 horses. At all the great
cities he had palaces, and the one at Bussorah was still to be
seen in the fourth century. His landed estate in Irâc was rated at
1,000 golden pieces a day. Abd al Rahmân had 1,000 camels,
10,000 sheep, and 100 horses, and he left property valued at
between three and four hundred thousand dinars. Zeid left gold
and silver in great ingots, and had land valued at 10,000 dinars.
The Coreishite nobles built themselves grand palaces in
Mecca and Medîna, and in their environs such as Jorf and Ackîck.
Othmân himself had a splendid palace at Medîna, with marble
pillars, walls of costly stucco, grand gates and gardens; he is also
said to have amassed vast treasures, though we are not told what
came of them after his overthrow.
Masûdy contrasts painfully all this luxury at home and abroad
with the frugal severity that prevailed even in the Caliphate of
Omar, who grudged to spend sixteen dinars on the pilgrimage to
Mecca.
[457] Quoting from the Corân (Sura ix. 36), where these words
are applied to Christian priests and monks; but Abu Dzarr gives
them here a more general application. See Life of Mahomet, p.
470; and Sprenger’s Leben des Mohammeds, vol. ii. p. cvi.
[458] Attempts are made by Abbasside tradition to show that
Abu Dzarr was driven into opposition by the tyranny of Muâvia’s
rule in Syria, and by divers ungodly practices at Medîna, which he
denounced as certain to bring down judgment on the city. But Ibn
al Athîr justly doubts this, and distinctly says that his preaching
tended to excite the poor against the rich. Abu Dzarr’s doctrines
were based on the equality of all believers; and the danger lay in
their popularity with the socialistic faction which decried the
pretensions of the Coreish. Before Muâvia, he reasoned thus:
‘Riches, ye say, are the Lord’s; and thereby ye frustrate the
people’s right therein; for the Lord hath given them to his people.’
‘Out upon thee!’ replied Muâvia; ‘what is this but a quibble of
words? Are we not all of us the Lord’s people, and the riches
belong unto the same?’ Tradition dwells on the poverty of Abu
Dzarr’s life at Rabadza to add point to Othmân’s unkind
treatment. The Beni Ghifâr, his tribe, are said to have resented his
ill-treatment by joining the insurgents when they appeared.
[459] On this subject historians say very little; and it is chiefly
from incidental notices in fragments of early poetry that Von
Kremer has so ably traced the inroads of profligacy and the
practice of forbidden pastimes—music, wine, and gambling. The
brief notice of Ibn al Athîr on this matter is as follows: ‘The
prevalence of a worldly spirit first showed itself at Medîna in the
flying of doves and shooting with pellets (with a gaming aim); and
in the eighth year of his Caliphate Othmân appointed an officer to
stop the same, who clipped the birds’ wings, and broke the cross-
bows.’ A citizen was rebuked by the Caliph for playing at
‘oranges’ (apparently some game of chance); and he thereupon
got angry and joined the hostile party. The anti-Omeyyad
tendency of the tradition on this subject is evident from Welîd (the
drunkard) being named as the person employed by the Caliph to
administer the rebuke.
[460] As enlarged by Othmân, the Mosque was 160 cubits
long, and 150 broad. As in Omar’s time, it had six gates for
entrance.
[461] Othmân defended his innovations as based on the
practice of the pilgrims from Yemen, who recited additional
prayers on behalf of their distant homes; and he too (he said) had
a property at Tâyif, as well as at Mecca. The matter seems at first
sight altogether insignificant. But in an established ritual, the
smaller the change, the greater oftentimes the scandal and
indignation, as we need not go far to see. And although no point
of doctrine was apparently involved, yet the practice of the
Prophet had come to be regarded as an obligatory precedent in
the commonest matters of daily life.
[462] For Abu Hodzeifa, see Life of Mahomet, p. 65. He left his
infant son to be brought up by Othmân, who faithfully discharged
the trust. When he grew up he asked for a government or military
command, but was told that he was not yet fit for it, and must
prove his capacity in the wars of Egypt and Africa. He never
forgot the slight, and was active in the insurgent ranks. Various
other examples are given of personal enmity, such as citizens
alienated by the reprimand for gaming, a chief imprisoned for the
ill-treatment of a Christian tribe, whose favourite hound he had
killed, and so on.
[463] The well was at the distance of two miles from the city.
Another well, called Rûma, was bought by Othmân, during the
Prophet’s lifetime, from the Jews for the use of the Moslems. He
first purchased the half title, the well being used day about by
either party; but on their alternate days the Mussulmans emptied
the well of enough water to last them two days. Whereupon the
Jewish owner insisted on Othmân’s purchasing the entire right,
which he did; and Mahomet promised him a fountain in Paradise
for the same.
[464] For traditions regarding Mahomet’s ring, see Life of
Mahomet, pp. 544 and 596. The despatches sent by him to the
several kings in the eighth year of the Hegira were attested by it.
The most received account is that the legend on it was ‘Mahomet,
Prophet of God’ (Mohammed Rasûl Allah, in three lines,
beginning from the bottom). It was used for all documents
requiring a seal, by Mahomet and his successors. The new ring
disappeared at the time of Othmân’s assassination, and, like the
original, was never seen again.
[465] One of the four wives who survived him was Omm al
Banîn, daughter of the famous freebooter, Oyeina. Othmân had
thirteen children, and (so far as we read) no issue by slave-girls,
which, looking to the habits of the time, is somewhat remarkable.
[466] The name of this demagogue was Abdallah ibn Saba,
but he was usually called Ibn Sauda, and was supposed to come
from Yemen. It is notable that this first sect of Alyites (if it can be
so called) was founded by a Jewish convert. What led him (if the
story of his teaching be not altogether a proleptic fiction of
tradition) to magnify Aly, who had hitherto put forth no claim, nor
indeed at any time dreamed of the extravagant pretensions in
store for him after his death, it is difficult to understand. Nor did
these transcendental notions regarding Aly gain any ground
whatever till a much later period. Ibn Sauda had evidently imbibed
some extreme notions on the dignity of prophets. ‘Strange,’ he is
reported to have said;—‘strange that men should believe in the
second coming of the Messiah, and not in that of Mahomet.’ The
idea, we are told, was inspired by the verse in Sura xxviii. v. 84,
‘Verily, he who hath given thee the Corân will surely bring thee
back again;’ which, of course, referred only to Mahomet’s
returning again to Mecca. Indeed, the whole account of this man’s
teaching is obscure and uncertain; and the Alyite notices of it may
be altogether anticipatory and unreal.
[467] The youth and his father belonged to the Beni Asad. On
hearing of the riot, Toleiha (the quondam prophet), chief of that
tribe, hastened with a body of his men to the palace for their
rescue; but found that both had escaped half dead.
Another version is, that on Saîd’s giving expression to the
sentiment about ‘the Sawâd being the Garden of the Coreish,’ the
whole company sprang to their feet and shouted excitedly: ‘Nay,
but the Lord hath given the Sawâd to us and to our swords.’ On
this, the captain of the body-guard retorted angrily at their rude
reception of his master’s words; whereupon they set upon him
and left him half dead. The inflammable material was all around,
and wanted only the spark to explode. This unfortunate speech
about ‘the Garden of the Coreish’ was in the mouths of the
disaffected all through the insurrection.
[468] The chief amongst them was Mâlik al Ashtar, of whom
we shall hear more as the most sanguinary amongst the traitors;
Zeid ibn Sohan; Jondob (already noticed); Orwa; and Thâbit ibn
Cays. Yezîd, a brother of the last, another chief leader of sedition,
was not sent. Muâvia wrote to Othmân that they were an ignorant
crew, bent on sedition, and on getting possession of the property
of the Zimmies, that is of the subject races, whose rights of
occupancy had been recognised as the hereditary tenants of the
Sawâd—a policy, as we have seen, firmly upheld by Omar
throughout Chaldæa, and which it was one object of the
malcontents to upset. According to one account, the exiles were
sent back by Muâvia, after expressing penitence, at once to Kûfa
—where, however, resuming their factious courses, Othmân, as a
last resource, despatched them again to Syria, this time to
Abdallah, Khâlid’s son, at Hims. Muâvia is throughout represented
as upholding the claims of the Coreish against the Arab faction,
showing thus the real aim of the ringleaders.
A story is told that the exiles, enraged at the menaces of
Muâvia, leaped upon him and seized him by the beard;
whereupon, shaking them off, he warned them that they knew
little of the loyal spirit of the Syrians, who, if they only saw what
they were doing, would be so enraged that it would be out of his
power to save their lives.
[469] Only two or three names are given of those who kept
aloof from seditious action: as Zeid ibn Thâbit (the collector of the
Corân); Hassân, the poet, his brother; Káb ibn Mâlik, and Abu
Oseid—all natives of Medîna; so that the whole body of Refugees
(the Coreish), excepting Othmân’s immediate kinsmen, must have
joined the treasonable faction.
[470] No doubt Aly spoke the truth. Yet Othmân’s weakness
towards the seditious populace was a far greater peril than his
tender treatment of his governors.
[471] I have given all this as I find it in tradition, but not without
some misgiving; especially of the part about Merwân, whom, as
the evil genius of Othmân, the Abbasside writers are never weary
of abusing.
[472] Amru, who had become a petulant malcontent ever
since his deposition, is represented as speaking contumeliously
of Othmân to his very face; and Othmân is represented as
returning it in kind, calling him ‘a louse in his garments.’ On one
occasion the Caliph is said to have addressed the people,
‘leaning on the staff of Mahomet’ (a venerable relic that had
descended from the Prophet to Abu Bekr and Omar), when an
Arab seized and broke it over Othmân’s head. Such stories,
however much they may be tinged with Abbasside exaggeration
and prejudice, point to the fact that Othmân was falling rapidly in
popular esteem.
[473] The four were Mohammed ibn Maslama, often employed
by Omar, as he had been by Mahomet himself, on confidential
missions; Osâma ibn Zeid, commander of the Syrian expedition at
Mahomet’s death; Abdallah, son of Omar; and Ammâr, whose
injudicious appointment by Omar to the governorship of Kûfa
appears to have turned his head, for he fell into the conspirators’
toils.
[474] We have abundance of conversations professing to have
passed between Othmân and his advisers; but they have no
further authority than as they represent the sentiments
conventionally attributed to the several speakers. As, however, it
may give point to the crisis now rapidly approaching, I subjoin the
following epitome of the most received account:—
Othmân: ‘Alas, alas! what is all this I hear of you, my deputies
and governors? I greatly fear that the complaints may be true;
and it is upon me the burden falleth.’ They replied that the Caliph
had sent his own men out to see, and they had found nothing
wrong. Then he asked what they advised him to do. Sád (ex-
governor of Kûfa) would have the traitors, who were burrowing in
the dark, unearthed and slain; then sedition would subside.
Muâvia: ‘In Syria there is no disaffection, and it would be
everywhere the same were the people fairly and firmly dealt with.’
Abu Sarh proposed to work through the Dewân, increasing or
diminishing stipends by way of reward and punishment. Ibn Aámir
advised to engage the restless spirits in war, and so the crisis
would pass over. Amru, embittered by his supersession in Egypt,
is represented as addressing Othmân in coarse abuse. Othmân
replied despondingly:—‘Cruel measures he would not sanction. If
rebellion was to come, no one should, at the least, have that to
say against him. The millstone would grind round and round to
the bitter end. Good had it been, if before it began to revolve, he
had been taken to his rest. There was nought left for him but to be
quiet and to see that no wrong was done to anyone.’ So he gave
the governors leave to depart, saying only that if fresh campaigns
were set on foot, he would approve of that; otherwise he would
hold on his way.
As they took their leave, Káb, the Jewish convert, said, ‘It will
be the grey mule that wins,’ meaning Muâvia, who overheard the
saying and from that moment (so the tradition runs) kept the
Caliphate in view.
Another scene is represented, in which Othmân, surrounded
by his own advisers, sends for Aly, Zobeir, and Talha. Muâvia
pleaded before them the cause of the aged Caliph, and warned
them of the danger they ran to their own selves in allowing any
attack calculated to abate the sacredness of the Caliph’s person;
it was, he said, both their duty and their interest to support him in
his feeble old age. On this, Aly reproached Muâvia as the son of
Hind, the ‘chewer of Hamza’s liver.’ ‘Let alone my mother,’ he
responded angrily; ‘she became a good believer, and after that
was not a whit behind thine own.’ Othmân interposed: ‘My cousin
Muâvia doth speak the truth. Now tell me wherein I have gone
astray, and I will amend my ways. It may be that I have been too
open-handed towards my kinsmen. Take back that which they
have received.’ So Abu Sarh disgorged 50,000 dirhems; and
Merwân 15,000, and they all departed for the moment satisfied.
But all these accounts must be received with suspicion. In the
midst of such violent factions as were springing up, the marvel is
that tradition has preserved so consistent a narrative as we have.
[475] Adapting the words from Sura xxxix. v. 39.
[476] For the Lesser Pilgrimage, or Omra, see Life of
Mahomet, p. xii. It may be performed in any month of the year, but
preferably in Rajab (three months earlier than the commonly
received date of the attack, which I have followed); and some
traditions accordingly give this as the date of the advance upon
Medîna. That, however, would make the interval (from January to
May) too long for the intervening events, which were hurried
through within the period of a couple of months, if so long.
[477] See above, p. 313.
[478] The men of Kûfa pitched at Al Awas; the Bussorah party
at Dzu Khashab; the Egyptians at Dzu Marwa—all places in the
close neighbourhood of the city.
[479] They marched off, we are told, expecting that the
citizens would break up their armed gathering as soon as they
were gone, and concerting to return again each from their
separate road.
[480] The history of the document is obscure. On the one
hand, it certainly was sealed with the Caliph’s signet; but who
affixed the seal, and whether it was surreptitiously obtained,
cannot be told. Nobody alleges Othmân’s complicity. Most
traditions attribute the writing and sealing of the order to Merwân,
the Caliph’s unpopular cousin, who, throughout the narrative,
receives constant abuse as the author of Othmân’s troubles; but
all this is manifestly tinged by Abbasside and anti-Omeyyad
prejudice. Aly’s objection of collusion between the three insurgent
bodies appears unanswerable. There must have been some
preconcerted scheme as to the simultaneous return of the three
camps; and there is a strong presumption of something unfair as
regards the document also. Amidst conflicting evidence, it is
beyond the historian’s power to offer any conclusive explanation.
It is, of course, possible that Merwân may have taken upon
himself the issue and despatch of the rescript; and, indeed, there
was not wanting ground for his venturing on such a course (and
something perhaps also to be said for his doing it unknown to
Othmân), excepting only the deception of the insurgents by false
promises. The insurgents may also have got scent of the
document before they started ostensibly with the purpose of
returning home. But these are all mere assumptions.
The Persian version of Tabari has a different story, namely,
that the Egyptian band, on seizing the document, turned their
faces back again towards Medîna, despatching at the same time
messengers to apprise the Kûfa and Bussorah bands of Othmân’s
treachery, and to recall them, so that all should reach Medîna and
join in the attack together. Neither Ibn al Athîr nor Ibn Khallicân
have anything to this effect, and it is hardly consistent with Aly’s
speech, noticed above. The Arabic original of Tabari, now being
published, may possibly throw further light on this chapter.
[481] Mohammed ibn Maslama, a Companion (as we have
seen) highly trusted both by Mahomet and his successors; and
Zeid ibn Thâbit, the collector of the Corân, tried to speak in
confirmation of what Othmân had said, but were violently silenced
and abused by the rebels Hakam ibn Jabala and Mohammed ibn
Coteira.
[482] There are traditions, but of an entirely Abbasside stamp,
of other interviews between Aly and the Caliph, with repeated
promises of the latter to amend; Aly recriminating that these
promises were no sooner made, even from the pulpit and before
the congregation, than under the baneful influence of Merwân
they were broken. Even Nâila, his wife, is represented as blaming
her weak-minded husband for his fickleness. But were all this
true, it would go but a little way to relieve Aly, Zobeir, and Talha
from the charge of desertion, or, worse, of treasonable collusion
with rebels against the rightful monarch—a short-sighted policy
even in their own interest.
[483] He is called Al Ghâficky, the ‘Ameer,’ or Commander of
the insurgents.
[484] According to some traditions, we are told, that Othmân
prevailed on Aly to procure for him a three days’ truce, under the
pretence of issuing orders to the governors for a reform of the
administration; and that he treacherously employed the time
instead in strengthening the defences, and excused himself by
saying that the time was too short to carry out the promised
reforms. But the story is altogether of the Abbasside type.
[485] The authorities are conflicting as to the length of the
siege, though the several stages of the attack and investment are
sufficiently well defined. After the first uproar Othmân still
presided at the daily prayers for thirty days, after which he was
besieged for forty days—that is ten weeks in all. Another tradition
is that after the blockade had lasted eight and forty days, tidings
of coming succour reached the city, and then the investment
became severe. But this would leave too long an interval—
namely, three weeks—between the report of help being on its way
and the final issue, before which the columns, hurrying from Syria
and Bussorah, should have had ample time to arrive at Medîna.
The Syrian column, we are told, reached as far as Wâdy al Cora,
and that from Bussorah as far as Rabadza, when they heard that
all was over, and accordingly turned back.
[486] The talk among the courtiers of Al Mâmûn, in the third
century, as reflected in the Apology of Al Kindy, was that Aly, even
at a much earlier period, contemplated the putting of Othmân to
death (Apology, p. 25). There seems to be no proof or
presumption of this; but anyhow, one cannot but feel indignant at
the attitude of Aly, who would do so much and no more; who sent
his son to join the Caliph’s guard at the palace gate, and was
scandalised at his being denied water to drink; and yet would not
so much as raise a finger to save his life.
We have also traditions in which Othmân is represented as
reproaching Talha for encouraging the insurgents to a more strict
enforcement of the blockade; but, whatever his demerits in
deserting the Caliph, this seems incredible. The ordinary account
is that Talha as well as Zobeir, on hearing of the rebel excesses,
kept to his house; others, again, say that they both quitted
Medîna.
Omm Habîba, as daughter of Abu Sofiân, naturally
sympathised with Othmân. Hantzala, one of the citizens of Kûfa
who had accompanied the insurgents, was so indignant at their
treatment of one of ‘the Mothers of the Faithful,’ that he went off to
his home, and there gave vent to his feelings in verses expressive
of his horror at the scenes his comrades were enacting at
Medîna.
One day, we are told, Othmân, goaded by the thirst of himself
and his household, ascended the roof, and cried aloud: ‘Ye men!
know ye that I bought the well Rûma, and furnished it with gear
that the Moslems might quench their thirst thereat? and now ye
will not let me have one drop to quench my thirst. Moreover, I
builded you such and such a mosque; and now ye hinder me from
going forth to say my prayers in the Great Mosque.’ And so on,
contrasting the various benefits he had conferred upon them, and
the kind and loving words the Prophet used to address to him,
with the cruel treatment he was now receiving; whereat the hearts
of all were softened, and the word was passed round to hold back
from pressing the attack. But Ashtar, the rebel, said, ‘He is but
playing with you and practising deceit,’ and so he resumed the
attack. There are many such traditions, but they seem to possess
little authority.
[487] The pilgrims, in order to reach Mecca in time for the
pilgrimage (beginning on the 8th of Dzul Hijj, June 7), must have
left Medîna a week or ten days previously; that is, some three
weeks before the final attack on the palace.
[488] The one killed was Moghîra, a Thackîfite from Tâyif. He
was a confederate of the Beni Zohra, the same who had
persuaded that clan to retire from the Coreishite army when it
marched forth to attack Mahomet at Bedr (Life of Mahomet, p.
228).
Merwân received a sword-cut, which severed one of the
tendons of the neck, and left him, when he recovered, with his
neck stiff and shortened. The rebels were about to despatch him
when his foster-mother cried out: ‘Do ye seek to kill him? he is
dead already; if ye would sport with and mutilate his body, that
were inhuman and unlawful.’ So they left him. In after days, when
Merwân came to power, he showed his gratitude to this woman
by giving her son a command.
[489] The blood, we are told, flowed down the leaves just
touching these words: ‘If they rebel, surely they are schismatics;
thy Lord will swiftly avenge you.’ (Sura ii. v. 138.) The
appropriateness of the text, however, may of itself have
suggested the story.
When the insurgents first rushed in, he was at the moment
reading the appropriate passage in Sura iii. v. 174. Referring to
the battle of Ohod, and the danger in which Medîna was then
placed, the disaffected citizens are there represented as taunting
Mahomet and his followers in these words: ‘Verily, the men (of
Mecca) have gathered forces against you; wherefore, be afraid of
the same. But (the taunt) only increased their faith, and they said:
The Lord sufficeth for us; He is the best Protector.’ This was a
favourite text of Othmân’s, and he may perhaps have turned to it
for comfort now that vain was the help of man.
[490] The actual murderers were Al Ghâficky, the leader, and
Sudân, who was himself killed. Kinâna ibn Bishr is also named.
All these belonged to the Egyptian band, which seems to have
contained the most rabid of the insurgents. Amr ibn al Hamac
leapt upon the body, hardly yet breathless, and inflicted nine
wounds—‘three for the Lord’s cause, and six to satisfy his own
passion.’
[491] These two were among the chief men ‘whose hearts
were gained over’ by largesses from the booty after the battle of
Honein.—Life of Mahomet, p. 436. Hakîm is frequently mentioned
in the Prophet’s biography. It was Hakîm who carried supplies to
his aunt Khadîja when shut up with Mahomet in the Sheb.—Life,
p. 100.
It is said that a party of the citizens of Medîna made an
attempt to stop the funeral, but desisted on seeing that a tumult
would arise. We are also told that Aly himself, on hearing of the
design to molest the procession and cast stones at the mourners,
did his best to prevent it. Indeed, Abbasside tradition abounds
with attempts to rescue the memory of Aly from the obloquy
attaching to the heartless part he had been acting. For example,
Masûdy gives us a tradition that when Aly heard that all was over
he hastened to the palace and asked his son how it had
happened—as though he could not for many days have foreseen
the fatal termination to which the blockade was tending!
[492] The field was called Hashh Kaukab—the Garden of the
Star.
[493] My impression, on the whole, is that it was an
afterthought. The narrative of those who side with Talha and
Zobeir is as follows: After Othmân’s death the city was for some
days in the hands of the insurgents. No one ventured to accept
the Caliphate. Sád and Zobeir had already quitted the city; and all
the members of the Omeyya clan who were able had effected
their escape to Mecca. The rebels themselves were at their wits’
ends: ‘If we quit Medîna,’ they said, ‘and no Caliph is appointed,
anarchy will burst forth everywhere. It appertaineth unto you
(addressing the men of Medîna) to appoint a Caliph. Wherefore
look ye out a man for the throne, and make him Caliph. We give
you one day’s grace for the same. If ye choose him, well; but if
not, then we shall slay Aly, Zobeir, and Talha, as well as a great
number of you.’ Alarmed at these threats, the leading citizens
repaired to Aly, who, at first, bade them seek another; but they
constrained him; and, as a last resource, to rid them of the
insurgents, he consented. Then they drew Zobeir (who, by this
time, had returned) and Talha to the Mosque, and forced them, at
the point of the sword, to swear.
It seems certain that the rebels of Kûfa and Bussorah were in
favour respectively of Zobeir and Talha; but that they were
induced to accept Aly, either through fear (as the partisans of the
two pretenders hold) of the Egyptian regicides, or because the
citizens made choice of Aly.
[494] Thus Sád, the conqueror of Irâc, refused to swear till all
else had done so; whereupon, Ashtar, head of the conspirators
from Kûfa, threatened to behead him; but Aly said, ‘Leave him
alone; I will be surety for him.’ Moghîra, also, and a company of
the late Caliph’s adherents, declined to swear, and were left
unmolested. Amongst them was Hassân, the poet, and his
brother Zeid (collector of the Corân), whom Othmân had
appointed Chancellor of the Exchequer. Of the latter, it is said that
when Othmân was first attacked, he cried to his fellow-citizens,
‘Ye men of Medîna, be ye Ansârs (Helpers) of the Lord for the
second time, even as ye were Ansârs of His Prophet at the first.’
But Abu Ayûb, another of the citizens, made answer and said,
‘Verily, he shall get no help from us. Let the multitude of his train-
band slaves be his Ansârs!’
[495] This servile population (Sabâya or ‘captives’) had been
pouring for years in a continuous stream, during the campaigns,
into Medîna. They were employed as domestic servants, warders,
body-guards, &c. Some followed trades, in quasi-freedom, paying
the profits to their masters. They mostly embraced the Moslem
faith because of the privileges it conferred. On the outbreak they
became insubordinate, and broke away into a defiant attitude.
This would occur the more readily as they formed the guards of
the treasury and mansions of the great men; and, being the only
trained force at Medîna, no doubt themselves felt their power. We
find them similarly taking part in the outbreak at Bussorah. Like
the Janissaries or Memlûk of later days, they were a petulant
brood. Immediately on homage being done to Aly, they are said to
have lampooned him in minatory verses, to which Aly (not to be
outdone by the poetry even of slaves) replied in extempore
couplets. Proclamation was made that slaves not returning to
their masters would be treated as outlaws, but it had no effect.
[496] The tradition runs that Moghîra, at the first, gave sound
and sincere advice to this effect; but that, finding Aly obstinately
opposed to it, he returned next day, saying that, on reflection, he
had changed his mind. When Ibn Abbâs came, Aly told him that
Moghîra had, at the first, attempted to deceive him, but on the
second day had spoken true, and advised him to put in his own
men. ‘Not so,’ said Ibn Abbâs; ‘just the reverse. It was the truth
which he spake at the first; the last was not his true opinion.’ And
so it turned out; for Moghîra, finding his advice disregarded,
departed to join the malcontents at Mecca
[497] Life of Mahomet, pp. 324 and 527.
[498] I have given this conference fully, because, in substance
at least, it shows the impracticable bent of Aly’s mind which
quickly drew on the civil war. It is also not unlikely in itself. The
purport of such a conversation would become known; and,
moreover, besides this and one or two other uncertain
conversations, we have little or nothing to explain the early events
of Aly’s Caliphate, and the motives which actuated him.
[499] Amru, it is said, pressed this course upon Muâvia,
saying, in his proverbial style, ‘Show the dam her foal, it will stir
her bowels.’
[500] The officers appointed were his cousin Abdallah ibn
Abbâs, a faithful adherent, and his brother Cutham; Omar, son of
Abu Salma (half-brother of Omm Salma, the Prophet’s widow);
Abu Leila, nephew of Abu Obeida; Aly’s own son Mohammed,
son of his Hanifite wife, &c.
[501] For this passage in the Prophet’s life see Life of
Mahomet, pp. 311 et seq.
[502] We are treading now on specially factious ground, and
have to weigh with care the bias of tradition which represents
Ayesha as suddenly converted from a deadly enemy of Othmân
into the champion of his memory. Thus, when, on receiving the
tidings of the murder on the way back from Mecca, she declared
that she would avenge his death: ‘What!’ cried her informant,
startled by her zeal; ‘is this thy speech now, whilst but yesterday
thou wast foremost to press the attack upon him as an apostate?’
‘Yea,’ she replied; ‘but even now he repented him of that which
they laid to his charge, and yet after that they slew him.’ In reply,
her informant recited these verses: ‘Thou wast the first to foment
the discontent. Thou commandedst us to slay the prince for his
apostasy, and now, &c.’ How far this has been invented (possibly
as a foil to Aly’s equally strange and inconsistent conduct), or
whether the inconsistency in Ayesha’s conduct was really as
strange as here represented, it is difficult to say. Anyhow, it must
be admitted that Ayesha was a jealous, violent, intriguing woman,
a character that may well account for much that would otherwise
appear strange.
[503] This famous camel is an object of special interest to
tradition. Some say it was bought for Ayesha in Yemen; others,
that it belonged to the Orni guide who piloted the expedition; and
that, in addition to a large sum for his services, he got the camel
purchased for Ayesha in exchange for his own.
[504] The women of Mecca accompanied Ayesha as far as
Dzât Ire. Some of Mahomet’s widows may have been at Mecca
just then for the pilgrimage, and, in the present troubled state of
Medîna, they may have preferred to stay on there. Perhaps some
of them may have settled permanently in the Holy City, On the
other hand, we know from a previous notice that Omm Habîba, at
any rate, still resided at Medîna.
[505] Saîd inquired of Talha and Zobeir which of them was to
be the Caliph. ‘Whichever,’ was the answer, ‘the people may
choose.’ ‘But,’ replied Saîd, ‘if ye go forth as the avengers of the
blood of Othmân, then the succession should of right devolve
upon his sons,’ two of whom were with the rebel force. ‘That,’ they
answered, ‘it will be for the chief men of Medîna eventually to
settle amongst them.’
[506] Meaning Talha and Zobeir themselves, and intimating
that these had had as much hand in fomenting the insurrection,
and were as responsible for its fatal result, as anybody else.
[507] Omm al Fadhl.
[508] The incident is adduced to show the alarm of Aly. He had
sent for Abdallah, Omar’s son, who declined to pledge himself to
join the army against the rebels of Bussorah till he saw what the
other citizens of Medîna did. On this he prepared to leave for
Mecca, assuring his stepmother (Aly’s daughter) that he meant to
keep aloof from the rebels, which he did.
[509] This column, which was got up in haste and with
difficulty (for there was no enthusiasm at any time for Aly), was
composed chiefly of men belonging to Kûfa and Bussorah in Aly’s
interest. What these were doing then at Medîna does not appear.
[510] The commander of the first campaign after Mahomet’s
death.
[511] Only one man of the band from Bussorah that attacked
Othmân, the warrior Horcûs, escaped, at the intercession of the
Beni Sád.
[512] Abu Mûsa, Governor of Kûfa, abused Ammâr, the envoy
of Aly, as a murderer. When urged by Hasan (whom on his arrival
he embraced affectionately) to support his father in putting down
the dissension that rent the people, Abu Mûsa replied that he had
heard the Prophet say that ‘in the event of sedition, walking was
better than riding, standing better than walking, and sitting better
than either.’ He exhorted the citizens, therefore, to adopt this
maxim, and, following the example of the Coreish, to sit still at
home;—‘if they studied their eternal interests, they would do this;
if only their temporal interests, they would go forth and fight.’ A
tumult arose; the palace was sacked, and he was deposed.
Hasan, spite of his want of ambition, must have managed the
business well, especially after the failure of the previous
deputations, which consisted of such able men as Mohammed,
son of Abu Bekr, the sons of Abbâs and Jafar, and Ashtar the
arch-regicide.
The spiritless rôle assigned by tradition to Hasan is illustrated
by a conversation which passed between him and his father at
Dzu Câr. Hasan: ‘Thou hast ever neglected my advice, my father,
and now thou wilt be deserted all round, and slain.’ Aly: ‘And thou
never ceasest whining like a girl. What advice of thine have I not
followed?’ Hasan replied that his father should have quitted
Medîna before Othmân was slain; after the murder, he should not
have accepted the Caliphate till the provinces had agreed in his
nomination; and now that Talha and Zobeir had risen up, he
should have stayed at home, and let them take the first offensive
step. To the first point Aly answered that at the time he was
himself besieged, and could not, even if he had so wished,
escape from Medîna; that he had been regularly elected, and
would fight it out to the end; that as for staying at home, he would
have been like a hyæna, baited by enemies on all sides; and that
if he did not look after his own interests, he saw no one else who
would do so for him. The conversation may be fictitious, but it
entirely accords with Hasan’s poor and unaspiring character.
[513] Mohammed son of Abu Bekr, the regicide, was with Aly
during the impending battle, which would seem to show that all
those concerned in the insurrection against Othmân were not kept
back. Possibly the order applied only to the Bedouins from Kûfa
that were so concerned.
[514] Among other things, Aly said to Zobeir: ‘Dost thou
remember the day when we both were with the Prophet among
the Beni Ghanam; and he looked on me and smiled, and I smiled
in return; and thou saidst to him, “Do not allow the son of Abu
Tâlib to vapour thus;” and he answered, “It is no vapouring to
quarrel about; thou doest him an injustice”?’ And Zobeir was
touched.
The attitude of Talha and Zobeir is variously represented.
They both appear to have assented to Aly’s proposals; and
(notwithstanding Talha’s speech about compulsory swearing of
allegiance) to have continued peaceful negotiations.
On the other hand, Abdallah son of Zobeir manifests the same
ambitious spirit which led him many years afterwards into
rebellion, and at one time nearly gained for him the Caliphate. He
is represented as now taunting his father with faint-heartedness in
swearing to Aly that he would not fight; and even persuading him
to release himself from the oath by the legal substitution of freeing
a slave.
Again, it is said that Zobeir was staggered when he heard that
Ammâr was in the field against him, in consequence of
Mahomet’s having once said that Ammâr would be slain by an
ungodly host (a matter of which we shall hear more below). The
general tenor of tradition is, that, from whatever cause, he retired,
without fighting, into the neighbouring valley, and there met his
death.
It is very difficult to weave a narrative at once faithful and
consistent out of all this. The conversations of the rebel leaders
with Aly must have been to a great extent conjectural; and the
surprise of both armies no doubt adds to the confusion of the
narrative as given by our authorities. The general outline,
however, is established.
[515] The Eastern traveller will recognise and appreciate the
illustration.
[516] This camel is a prominent subject in tradition, as we
might expect from its having given its name to the battle, and
many tales of heroism are told both in its attacks and defence.
One says he never heard anything so fearful as the scream it
gave when hamstrung.
[517] The numbers may be exaggerated; but the loss of life
was, no doubt, immense, and it is evidence of the terrible fury
with which the battle was fought. Of one tribe, the Beni Dhabba,
alone, 1,000 men are said to have been slain. The strong partisan
feelings both of Bedouin against Coreish, and of the opposing
families of Hâshim and Omeyya, long pent up, tended to give
bitterness to the conflict; and there was in addition the new cry of
vengeance for the blood of Othmân.
[518] So carefully were Aly’s orders against plundering
observed, that whatever was found on the field, or in the
insurgent camp, was gathered together in the Great Mosque, and
every man was allowed to claim his own. To the malcontents who
complained that they were not allowed to take booty, Aly replied
that the rights of war, in this case, lasted only so long as the ranks
were arrayed against each other; and that, immediately on
submission, the insurgents resumed their rights and privileges as
brother Moslems.
[519] ‘She of the two shreds.’—Life of Mahomet, p. 145.
[520] There is a great abundance of tradition concerning
Ayesha, both in the battle and after it. In the heat of the action,
Aly’s soldiers taunted her as ‘the unnatural Mother of the Faithful.’
The soldiers on her side, in reply, extemporised a couplet,
extolling her as ‘the noblest and best of Mothers.’ When they told
it to her, she was much affected, and exclaimed, ‘Would that I had
died twenty years before this!’ Aly also, when he heard it, said,
‘Would that I too had died twenty years ago!’
Ayesha, always ready in repartee, was not very particular in
her language, and some of the speeches attributed to her are
both coarse and intemperate. Asim approaching her litter on the
field, she cursed him for the liberty he had taken. ‘It was but a
little something red and white,’ he said, impudently, ‘that I caught
a glimpse of.’ ‘The Lord uncover thy nakedness,’ she cried
angrily; ‘cut off thy hands, and make thy wife a widow!’ All which,
they say, came to pass. A saucy passage is related between her
and the aged Ammâr, whose last words were, as she was
leaving, ‘Praise be to the Lord that we shall hear no more that vile
tongue of thine!’
Aly’s conduct was forbearing and generous. Of the family with
which Ayesha was lodged at Bussorah, two sons had been killed
fighting, one on the side of Aly, the other against him. The widow
of the latter was loud in her lamentation, crying out against Aly as
the cause of her sorrow. Aly was asked to punish her; but he
refused, saying she was but a weak woman, and should not be
touched. On the other hand, some one who spoke contumeliously
of Ayesha was, by his order, beaten with shoes.
As Ayesha was starting for Mecca, Aly and a company
gathered round her. When the time came to bid farewell, she said,
‘Let us not entertain hard thoughts one against the other; for
verily, as regardeth Aly and myself, there happened not anything
between us (alluding to her misadventure in the Prophet’s
lifetime[521]) but that which is wont to happen between a wife and
her husband’s family; and verily Aly was one of the best of them
that entertained suspicions against me.’ Aly replied: ‘She
speaketh the truth; there was nought, beyond what she saith,
between her and me.’ And then he went on to say (quoting
Mahomet’s own words) that ‘she was not only the Prophet’s wife
in this world, but equally his spouse in the next.’
[521] Life of Mahomet, p. 311.
[522] A separate chapter is generally assigned by the Arabian
historians to this episode; but its interest lies almost wholly in the
intense hatred conceived by the usurper towards Othmân.
Hearing of his factious courses, Othmân, to soften and remind
him of his past care and favour, sent him from Medîna a purse,
and also a camel laden with rich garments, as a present. The
ungrateful rebel hung these up in the Great Mosque of Fostât,
and used them to point his invectives against Othmân and the
corruption of the age. Having joined the insurgent faction, he, no
doubt, hoped that Othmân’s successor would have confirmed him
in the government of Egypt. But Aly, treating him as he deserved,
showed him no favour, and appointed a man of his own to the
government.

You might also like