Game-Theoretic Learning and Distributed Optimization in Memoryless Multi-Agent Systems
This book presents new efficient methods for optimization in realistic large-scale, multi-agent systems. These methods do not require the agents to have the full information about the system, but instead allow them to make their local decisions based only
- PDF / 3,901,981 Bytes
- 176 Pages / 439.42 x 683.15 pts Page_size
- 60 Downloads / 148 Views
Game-Theoretic Learning and Distributed Optimization in Memoryless Multi-Agent Systems
Game-Theoretic Learning and Distributed Optimization in Memoryless Multi-Agent Systems
Tatiana Tatarenko
Game-Theoretic Learning and Distributed Optimization in Memoryless Multi-Agent Systems
123
Tatiana Tatarenko TU Darmstadt Darmstadt, Germany
ISBN 978-3-319-65478-2 DOI 10.1007/978-3-319-65479-9
ISBN 978-3-319-65479-9 (eBook)
Library of Congress Control Number: 2017952051 © Springer International Publishing AG 2017 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
Abstract
Learning in potential games and consensus-based distributed optimization represent the main focus of the work. The analysis of potential games is motivated by the game-theoretic design, which renders an optimization problem in a multi-agent system a problem of potential function maximization in a modeled potential game. The interest to distributed consensus-based optimization is supported by growing popularity of dealing with networked systems in different engineering applications. This book investigates the algorithms that enable agents in a system converging to some optimal state. These algorithms can be classified according to information structures of systems. A common feature of the procedures under consideration is that they do not require agents to have memory to follow the prescribed rules. A general learning dynamics applicable to memoryless systems with discrete states and oracle-based information is presented. Some settings guaranteeing an efficient behavior of this algorithm are provided. A special type of such efficient general learning procedure, called logit dyn
Data Loading...
