Full download Multiagent Systems: Introduction and Coordination Control 1st Edition Magdi S. Mahmoud pdf docx
Full download Multiagent Systems: Introduction and Coordination Control 1st Edition Magdi S. Mahmoud pdf docx
com
https://textbookfull.com/product/multiagent-
systems-introduction-and-coordination-control-1st-
edition-magdi-s-mahmoud/
DOWLOAD NOW
https://textbookfull.com/product/management-control-with-
integrated-planning-models-and-implementation-for-sustainable-
coordination-lukas-rieder/
https://textbookfull.com/product/coordination-cooperation-and-
control-the-evolution-of-economic-and-political-power-randall-g-
holcombe/
https://textbookfull.com/product/power-systems-control-and-
reliability-electric-power-design-and-enhancement-1st-edition-
isa-s-qamber/
https://textbookfull.com/product/ultimate-rust-for-systems-
programming-master-core-programming-for-architecting-secure-and-
reliable-software-systems-1st-edition-mahmoud-harmouch/
The Coordination of European Public Hospital Systems:
Interests, Cultures and Resistance 1st Edition Sorin
Dan (Auth.)
https://textbookfull.com/product/the-coordination-of-european-
public-hospital-systems-interests-cultures-and-resistance-1st-
edition-sorin-dan-auth/
https://textbookfull.com/product/natural-fiber-textile-composite-
engineering-1st-edition-magdi-el-messiry/
https://textbookfull.com/product/protective-armor-engineering-
design-1st-edition-magdi-el-messiry-author/
https://textbookfull.com/product/autonomous-agents-and-
multiagent-systems-aamas-2016-workshops-best-papers-singapore-
singapore-may-9-10-2016-revised-selected-papers-1st-edition-
nardine-osman/
https://textbookfull.com/product/introduction-to-process-control-
analysis-mathematical-modeling-control-and-optimization-1st-
edition-victor-a-skormin-auth/
Multiagent Systems
Introduction and
Coordination Control
Magdi S Mahmoud
Distinguished Professor, Systems Engineering Department
King Fahd University of Petroleum and Minerals
Dhahran, Saudi Arabia
p,
p,
A SCIENCE PUBLISHERS BOOK
A SCIENCE PUBLISHERS BOOK
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted,
or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, includ
ing photocopying, microfilming, and recording, or in any information storage or retrieval system, without written
permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copyright.com
(http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers,
MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety
of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment
has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identification and explanation without intent to infringe.
Multiagent system (MAS) is perhaps one of the most exciting and fastest growing
domains in the intelligent resource management and agent oriented technology which
deals with modeling of autonomous decision making entities. The field of multiagent
dynamic systems is an inter-disciplinary research field that has become very popular
in recent years, in parallel with the significant interest in the practical applications of
such systems in various areas, including robotics.
Recent developments have produced very encouraging results in its novel
approach to handle multi-player interactive systems. In particular, the multiagent
system approach is adapted to model, control, manage or test the operations and
management of several system applications, including multi-vehicles, microgrids
and multi-robots, where agents represent individual entities in the network. Each
participant is modeled as an autonomous participant with independent strategies
and responses to outcomes. They are able to operate autonomously and interact
proactively with their environment.
In recent publication, the problem of information consensus is addressed, where
a team of vehicles must communicate with each other in order to agree on key pieces of
information that enable them to work together in a coordinated fashion. The problem
is particularly challenging because communication channels have limited range and
experience fading and dropout. Along a parallel avenue, various topics regarding
synchronization and consensus in multiagent systems were examined. The results
demonstrated that the joint presentation of synchronization and consensus allows the
reader to learn about the similarities and differences of both concepts. Cooperative
control of multiagent dynamical systems interconnected by a communication
network topology was also studied. Using the terminology of cooperative control,
each system is endowed with its own state variable and dynamics. A fundamental
problem in multiagent dynamical systems on networks is the design of distributed
protocols that guarantee consensus or synchronization, in the sense that the states of
all the systems reach the same value.
Preface • v
In view of the available results, it turns out that research avenues in multiagent
systems offer great opportunities for further developments from theoretical, simulation
and implementations standpoints. This volume provides ‘‘system dynamics and
control perspective” of multiagent systems, with focus on mathematical modeling
of multiagent systems and paying particular attention to the agent dynamics models
available in the literature. We provide a number of problems on coordination and
control of multiagent systems which have gained significant attention recently as
well as various approaches to these problems. Looked at in this light, it has the
following objectives:
1. It gathers together the theoretical preliminaries and fundamental issues related
to multiagent systems.
2. It provides coherent results on adopting multiagent framework for critically
examining problems in smart microgrid systems.
3. It presents advanced analysis of multiagent systems under cyber-physical attacks
and develops resilient control strategies in order to guarantee safe operation.
October 2019 Magdi S Mahmoud
Acknowledgement
In writing this volume, I took the approach of referring within the text to papers
and/or books which I believe taught me some concepts, ideas and methods. I further
complemented this by adding some remarks and notes within and at the end of each
chapter to shed some light on other related results.
A particular manifestation of our life is due the ‘‘interaction” with people
in general and colleagues and friends. This is true in my technical career, where
I benefited from listening, discussing and collaborating with several colleagues.
Foremost, I would like to express my deep gratitude to Prof Yuanqing Xia, BIT,
China, who has been a good supporter, but most importantly an example of a true
friend. I owe him for his deep insight as well as for his trust and perseverance.
The process of fine tuning and producing the final draft was pursued at the
Distributed Control Research Group (DCRG), Systems Engineering Department and
special thanks must go to my colleague Dr Nezar M Alyazidi and graduate students
Mojeed O Oyedeji and Bilal J Karaki for helpful comments, superb interactions
and assistance throughout the writing of this book. The great effort by Dr Mutaz M
Hamadan was instrumental in producing the last version of the book. Portions of this
volume were developed and upgraded while offering the graduate courses SCE-612
171, SCE-612-172, SCE-701-172, SCE-701-181, SCE-515-182 at KFUPM, Saudi
Arabia.
It is a great pleasure to acknowledge the financial funding afforded by the
deanship of scientific research (DSR) through project no. BW 181006 and for
providing a superb competitive environment and overall support of research activities
at KFUPM.
I would appreciate any comments, questions, criticisms or corrections that
readers may kindly provide to me at my emails:
[email protected] or [email protected].
Contents
Preface iv
Acknowledgement vi
Author Biography xiii
1. Introduction 1
1.1 Overview 1
1.2 Elements of Graph Theory 2
1.2.1 Basic results 3
1.2.2 Laplacian spectrum of graphs 3
1.2.3 Properties of adjacency matrix 4
1.2.4 Nonlinear stochastic dynamical systems 7
1.2.5 Complex dynamical systems 12
1.2.6 Delay effects 15
1.2.7 Sampled-data framework 16
1.3 Multiagent System Approach 17
1.3.1 Practical examples 17
1.3.2 Some relevant definitions 18
1.4 Mathematical Models for Agent Dynamics 20
1.4.1 Single integrator model 21
1.4.2 Double integrator model 21
1.4.3 Uncertain fully actuated model 22
1.4.4 Non-holonomic unicycle model 23
1.5 Coordination and Control Problems 23
1.5.1 Aggregation and social foraging 24
1.5.2 Flocking and rendezvous 25
1.5.3 Synchronization of coupled nonlinear oscillators 25
1.6 Scope and Book Layout 27
viii • Multiagent Systems: Introduction and Coordination Control
2. Theoretical Background 29
2.1 Preliminaries of Distributed Systems 29
2.1.1 Problem description 30
2.1.2 Control design scheme 32
2.1.3 Without communication delays 33
2.1.4 With communication delays 37
2.2 Networked Multiagent Systems 38
2.2.1 Consensus in networks 40
2.2.2 The f-consensus problem 41
2.2.3 Iterative consensus and Markov chains 42
2.3 Applications 42
2.3.1 Synchronization of coupled oscillators 42
2.3.2 Flocking theory 43
2.3.3 Fast consensus in small-worlds 43
2.3.4 Rendezvous in space 44
2.3.5 Distributed sensor fusion in sensor networks 44
2.3.6 Distributed formation control 44
2.4 Information Consensus 45
2.4.1 Algebraic connectivity and spectral properties 47
2.4.2 Convergence analysis for directed networks 47
2.4.3 Consensus in discrete-time 50
2.4.4 Performance of consensus algorithms 52
2.4.5 Alternative forms of consensus algorithms 54
2.4.6 Weighted-average consensus 57
2.4.7 Consensus under communication time-delays 58
2.5 Consensus in Switching Networks 59
2.6 Cooperation in Networked Control Systems 61
2.6.1 Collective dynamics of multivehicle formation 62
2.6.2 Stability of relative dynamics 63
2.7 Simulation Studies 64
2.7.1 Consensus in complex networks 64
2.7.2 Multivehicle formation control 66
2.8 Notes 67
Introduction
1.1 Overview
The field of coordinated multiagent dynamic systems, including swarms and
swarm robotics, is a relatively new field that has become popular in recent years.
Since the pioneering work [1] on simulation of a flock of birds in flight using
a behavioral model based on a few simple rules and only local interactions, the
field has witnessed many developments. Currently, there is significant interest in
the applications of the field in various areas involving teams of manned or un
manned aerial, ground, space or underwater vehicles, robots and mobile sensors,
to name a few [2]–[9].
Because of the interdisciplinary nature of the field, the literature on coordi
nated multiagent dynamic systems have a moderately wide spectrum of perspec
tives. This chapter focuses on the system dynamics and control perspective with
the aim of presenting a short review on mathematical modeling, coordination and
control of multiagent dynamical systems.
Integrator and double integrator models are the simplest abstraction, upon
which a large part of results on consensus of multiagent systems have been
based, see [34], [35], [36], [41], [42], [43]. To deal with more complex mod
els, a number of recent papers are devoted to consensus of multiple LTI systems
[37], [38], [39], [40], [44], [45], [46], [47], [48], [49], [50], [51]. These results
keep most of the concepts provided by earlier developments, and provide new
design and analysis techniques, such as LQR approach, low gain approach, H∞
approach, parametrization and geometric approach, output regulation approach,
and homotopy-based approach. However, most of these results [37], [38], [39],
[40], [44], [46], [50], [51] mainly focus on fixed interaction topology, rather
than time-varying topology. How do the switches of the interaction topology and
2 � Multiagent Systems: Introduction and Coordination Control
agent dynamics jointly affect the collective behavior of the multiagent system?
Attempts to understand this issue have been hampered by the lack of suitable
analysis tools. The results of Scardovi et al. [45] and Ni et al. [40] are men
tioned here, because of their contributions to dealing with switching topology
in the setup of high-order agent model. However, when dealing with switching
topology, [45] and [40] assumed that the system of each agent is neutrally sta
ble; thus, it has no positive real parts eigenvalues. This assumption was widely
assumed in the literature when the interaction topology is fixed or switching.
Unfortunately, when the agent is stabilizable and detectable rather than neutrally
stable, and when the interaction topology is switching, there is no result reported
in the literature to investigate the consensus of these agents.
L associated with the zero eigenvalue. The set of neighbors of node si is denoted
by Ni = {s j ∈ V : (si , s j ) ∈ E}. A directed path is a sequence of ordered edges
of the form (si1 , si2 ), (si2 , si3 ), . . . , where si j ∈ V in a directed graph. A directed
graph is said to be strongly connected, if there is a directed path from every node
to every other node. Moreover, a directed graph is said to have spanning trees, if
there exists a node such that there is a directed path from every other node to this
node.
Let Re(z),Im(z) and �z� be the real part, the imaginary part and the modulus
of a complex number z, respectively. Let In (0n ) be the identity (zero) � matrix of
dimension n and 1n be the n × 1 column vector of all ones. Here, represents
the Kronecker product.
A graph G is defined as
G = (V, A) (1.3)
where V is the set of nodes (or vertices) V = {1, . . . , N} and A ⊆ V × V the set
of edges (i, j) with i ∈ V, j ∈ V. The degree d j of a graph vertex j is the number
of edges which start from j. Let dmax (G) denote the maximum vertex degree of
the graph G.
where D(G) is the diagonal matrix of vertex degrees di (also called the valence
matrix). Eigenvalues of Laplacian matrices have been widely studied by graph
theorists. Their properties are strongly related to the structural properties of their
associated graphs. Every Laplacian matrix is a singular matrix. By Gershgorin
theorem [15], the real part of each nonzero eigenvalue of L(G) is strictly positive.
For undirected graphs, L(G) is a symmetric, positive, semidefinite matrix, that
only has real eigenvalues. Let S(L(G)) = {λ1 (L(G)), . . . , λN (L(G))} be the spec
trum of the Laplacian matrix L associated with an undirected graph G arranged
in non-decreasing semi-order. Then,
• Property 3:
1. λ1 (L(G)) = 0 with corresponding eigenvector of all ones, and λ2 (L(G)) iff
G is connected. In fact, the multiplicity of 0 as an eigenvalue of L(G) is
equal to the number of connected components of G.
2. The modulus of λi (L(G)), i = 1, . . . , N is less then N.
The second smallest Laplacian eigenvalue λ2 (L(G)) of graphs is probably the
most important information contained in the spectrum of a graph. This eigen
value, called the algebraic connectivity of the graph, is related to several impor
tant graph invariants, and it has been extensively investigated.
Introduction � 5
Let L(G) be the Laplacian of a graph G with N vertices and with maximal
vertex degree dmax (G). Then properties of λ2 (L(G)) include
• Property 4:
1. λ2 (L(G)) ≤ (N/(N − 1)) min{d(v), v ∈ V};
2. λ2 (L(G)) ≤ v(G) ≤ η(G);
3. λ2 (L(G)) ≥ 2η(G)(1 − cos(π/N));
4. λ2 (L(G)) ≥ 2(cos Nπ − cos 2 Nπ )η(G) − 2 cos Nπ (1 − cos Nπ )dmax (G)
where v(G) is the vertex connectivity of the graph G (the size of a smallest set of
vertices whose removal renders G disconnected) and η(G) is the edge connectiv
ity of the graph G (the size of a smallest set of edges whose removal renders G
disconnected) [17].
Further relationships between the graph topology and Laplacian eigenvalue
locations are discussed in [14] for undirected graphs. Spectral characterization
of Laplacian matrices for directed graphs can be found in [15], see also Fig. 1.1.
A lemma about Laplacian L associated with a balanced digraph G is given
hereafter:
Lemma 1.3
If G is balanced, then there exists a unitary matrix
⎡ 1
∗ ... ∗
⎤
√
n
⎢ √1 ∗ . . . ∗ ⎥
⎢ n ⎥ m×n
V =⎢ ⎢ .. .. .. ⎥ ∈ C (1.5)
⎥
⎣ . . . ⎦
√1 ∗ . . . ∗
n
such that
� �
0
V ∗ LV = = Λ ∈ Cn×n , H ∈ C(n−1)×(n−1) (1.6)
H
"I don't believe she said so," said my mother; and she
translated Deborah's remark for Jeanne's benefit.
"What did you say to Mrs. Betty then, when she asked
you about it?" demanded Deborah, beginning to calm down
a little.
"I see how it is," said my aunt, who with all her
easiness of disposition was not a person to be despised.
"Deborah has allowed herself to be prejudiced, and to
believe her mistress capable of the most unworthy
conduct."
"I own it, madame," sobbed Jeanne in her turn; "I have
been too hasty; but to be called a Jesuit, when I have
suffered so much by them; and then my beautiful salad,
which the young master used to like so much in France—"
"But what?"
"Where does she keep her fiery eyes, if she has no head
—in her pocket?" I asked, laughing at this very original
ghost. "Perhaps she carries them on a dish before her, like
St. What's-her-name in the picture."
Aunt Amy had had time to cool. "Ah, well, I dare say it
was but thoughtlessness; and young maids must be young
maids, I suppose."
"Dear aunt, I will tell you how it was," said I. "We were
all gathering lavender-flowers for the still, and I saw that
Peggy, the still-room maid, had been crying, and asked
what was the matter. She said the mistress had been
scolding her because she had on ragged stockings, and
because she did not keep her head neat; and Betty asked
me if I did not think that was hard on the poor girl, when
she had so much to, do. And I said no: if I were her
mistress I would make her knit her own hose and wear a
clear-starched cap every day, as the maids do in Normandy.
Then Meg laughed, and said I would be a pattern
housekeeper, no doubt; and I said I did not believe I should
ever be as good-natured as you were. That was the whole
of it. I am sure nothing was farther from my thoughts than
any disrespect; and as to your housekeeping, I think it is as
perfect as can be—only, of course, many of the ways are
different from ours, and when I notice them 'tis natural to
speak of them."
"And what did you say?" asked my mother. "I said, 'Not
to leave you;' and besides, since I had come down here and
learned to know the people, I liked the place; and so I do.
Only I shall not like it, I am sure, if my aunt turns against
me."
"Yes, and you went and told your mother, and she
lectured me this morning about turning your head with
stories of London tine gentlemen."
"Did you not want me to tell, then?" said she. "I never
thought of that. I have no secrets from my mother."
"I shall leave them in your care," said he; "only old
Joslyns must take them out now and then or they will forget
how to fly."
"I do not mean to mind her," said I, "if only she would
not make mischief. But I think it is too bad in her to lead
me on to tell her about London and my uncle and aunt
there, and then go and tell your mother, as if it had been all
my doing. And then—but there, what is the use?" I added.
"You cannot understand, and there is no need of troubling
you with the matter. Only I wish we had stayed in Jersey—
that is all," I concluded, with a quiver in my voice.
"Yes indeed; but do you not want more than one pair?"
"I did not suppose you would have time for more than
one."
"So you ought to be. The knights of old could forge their
own armor, you know. But I think you are a little of
everything," said I. "If ever we should be cast away upon a
desert island, like the folks you read of yesterday, you could
set up housekeeping, and make yourself a great king
among the people."
"Only the woman who cares for it; and she is worth
seeing too," answered Andrew. "Is not this the old dame's
cottage?"
"I have them, but I fear they are too dry to grow," said
I. "However, it can do no harm to try."
(Two of them did grow, and are now fine bearing trees.)
"Oh, yes, and write as well. The people who took her in
were of the better class. They were not Cornish folk, but
East Country English, who came and settled here in the
reign of Charles the First. No one knew much about them,
and I fancy they might have had their own reasons for
keeping quiet, but my father never would allow them to be
molested. See, here is the cabinet I made from the old
chestnut-tree."
"Oh, no; I can walk from the Well House as well as from
here."
"But the way is very lonely, and you must pass the
Pisky Bank going and coming," said Margaret. "Won't you
be afraid?"
"I did not see his face, but he was tall and slim, with a
fair love-lock, which slipped out from under his cloak. That
was all I noticed, but somehow, he made me think of young
Mr. Lovel."
CHAPTER XIII.
THE BOOK.
"Have you not read it?" she asked, seizing and opening
it. "It looks delightful."
"No, I have not read it, and shall not till I show it to
maman," I answered.