Multiagent Systems: Algorithmic, Game-Theoretic, and Logical Foundations

Game theory is a personal favourite. I used a different book when I studied a course in game theory. That book served me but my sustained interest in the field made me look out for reputable game theorists and their publications. I am excited I came across this book. It gives an algorithmic perspective to the concepts it presents. Unlike other book, the book addresses game theory from the true lenses of computer science. If you are interested in algorithms and their application to games this book is a must read.

I used this book to learn more about auctions, mechanism design and the basics of game theory. Those ideas are becoming more and more popular in the crypto community but also the multi agent reinforcement learning community. The breadth of the book is massive but I did appreciate how most of the chapters can be read somewhat independently, makes it easier to pick up and record core concepts you wanna learn more about.

What did I like? Well, the book is well structured and helps with deeper learning of theory of MAS (for example, because of not just mentioning of a book/article that deals deeper with this or that problem, but even giving a kind of very brief review of it). What else? It is very nice that the authors do not try to seem "too clever" and are able to explain even rather difficult themes in such a way that even a person for whom English is not a mother tongue (I mean myself) reads this book without too big problems with understanding. And what I also liked a lot is a very long bibliography (13 pages). What did I dislike? I think I will be able to answer this question in several months, after I have given a course of lectures based on this book. Well, let's see... Now I have read the whole book and I do not see any serious disadvantages.

I have been looking for books on game theory and software development ever since I completed my game theory class. I am still reading this book, but I really like it so far.

This book was very useful to me. It teach me everything we need to know about Multiagent systems, as game theory.

Good book

A good academic text book, however, extremely theoretical and with some surprising "voids" on important topics. The book had served its purpose better if a few more examples and references to industrial cases were included. Some explanations fall short of what I suspect to be their intention at times. Students using this book would enjoy it as a book of reference for development. But as an aid to help them place MA in a business perspective it fails.

GREAT PRODUCT!

Ich habe mir mehr ein Buch in Richtung praktische Realisierung von Multiagent-Systems erwartet. Stattdessen ist es eine umfassende aber gleichzeitig nirgends sehr tiefgehende Sammlung von allen möglichen Mathematischen Resultaten. Wobei man in der Regel nie das Gefühl hat, dass das zu irgendwas gut ist. Z.B. wird bewiesen, dass unter bestimmten - nicht sehr realistischen Annahmen - eine Auktion, bei der der Sieger das Gebot des Zweitplazierten bezahlen muss, fair und optimal ist. Ich kenne keine derartige Auktion. Das Theoretische Grundproblem ist: Die meisten Sätze hängen direkt oder indirekt vom Konzept des Nash-Equilibriums ab. Mathematisch ist der Satz eine relativ triviale Folgerung des Brouwerschen Fixpunktsatzes. John Nash rechnet es selbst nicht zu seinen wichtigsten Mathematischen Resultaten. Aber die Ökonomen lieben den Satz, weil er so schöne Marktgleichgewichte ergibt. Die man allerdings in der Praxis nicht beobachtet. Man beobachtet es auch nicht in anderen Spielsituationen und selbst wenn man etwa für ein Programm die Equilibrium Strategie berechnen kann, dann spielt das Programm nicht besonders gut. Es wird bestensfalls ein schwer besiegbarer Remisschieber. Es versucht nicht die Fehler des Gegners auszunützen. Ein ähnlich gelagerts Buch ist Noam Nisan et al.: Algorithmic Game Theory. Teilweise überschneiden sich die Kapitel direkt. Es gilt auch für dieses Buch das oben Gesagte.

Very good read and quite easy to understand if you are into math/computer science. It mainly focuses on the more formal definitions, proofs etc. but also has quite a lot of discussion material as well. It is a bit vague in some points like the difference of pure and mixed strategy Nash equilibria and their existence, but I think the author makes their point clearly understood in the vast majority of the material covered.