With the advent of extremely affordable computing power, the world is becoming filled with distributed systems of computationally sophisticated components. However, no current scientific discipline offers a thorough understanding of the relation of such "collectives" and how well they meet performance criteria.
"Collectives and Design of Complex Systems" lays the foundation for the study of collective intelligence and how these entities can be developed to yield optimal performance. Partone describes how some information-processing problems can only be solved by the joint actions of large communities of computers, each running their own complex, decentralized machine-learning algorithms. Part two offers general analysis on the dynamics and structures of collectives. Finally, part three addresses economic,
model-free, and control-theory approaches to designing these complex systems. The work assumes a modest understanding of basic statistics and calculus.
Using an approach that integrates key theoretical principles with applications in real-world scenarios, this unique monograph surveys the latest research on the dynamics of collectives, their A.I. aspects, and critical design issues pertaining to them. Computer scientists, computer engineers, and all practitioners, researchers, and graduate students with an interest in this new and growing field will find the book an authoritative introduction and resource.
Increasingly powerful computers are making possible distributed systems comprised of many adaptive and self-motivated computational agents. Such systems, when distinguished by system-level performance criteria, are known as "collectives."
Collectives and the Design of Complex Systems lays the foundation for a science of collectives and describes how to design them for optimal performance. An introductory survey chapter is followed by descriptions of information-processing problems that can only be solved by the joint actions of large communities of computers, each running its own complex, decentralized machine-learning algorithm. Subsequent chapters analyze the dynamics and structures of collectives, as well as address economic, model-free, and control-theory approaches to designing complex systems. The work assumes a modest understanding of basic statistics and calculus.
Topics and Features:
Introduces the burgeoning science of collectives and its practical applications in a single useful volume
Combines approaches from known researchers in the physics, economics, game theory, biology, and artificial intelligence (A.I.) communities
Analyzes the role of information transmission in collectives, such as in the minority game, specific game theory scenarios, and computational ecologies
Presents a broad view of state-of-the-art techniques and developments in the field
Describes the benefits and promise of collectives-based design of systems, and explores recent challenges faced by researchers
Highlights the interdisciplinary value to studying and using collectives
Provides a comprehensive subject index and chapter-ending bibliographies
With an accessible style that integrates key theoretical principles with applications in real-world scenarios, this unique monograph surveys the latest research on the dynamics of collectives, their A.I.-related aspects, and critical design issues pertaining to them. Computer scientists, computer engineers, and practitioners, researchers, and graduate students with an interest in this new and growing field will find the book an authoritative introduction and resource.
