Optimal Payo Functions for Members of Collectives

نویسندگان

  • David H. Wolpert
  • Kagan Tumer
چکیده

We consider the problem of designing perhaps massively distributed collectives of computational processes to maximize a provided world utility function We consider this problem when the behavior of each process in the collective can be cast as striving to maximize its own payo utility function For such cases the central design issue is how to initialize update those payo utility functions of the individual processes so as to induce behavior of the entire collective having good values of the world utility Tradi tional team game approaches to this problem simply assign to each process the world utility as its payo utility function In previous work we used the Collective Intelli gence COIN framework to derive a better choice of payo utility functions one that results in world utility performance up to orders of magnitude superior to that ensuing from use of the team game utility In this paper we extend these results using a novel mathematical framework We review the derivation under that new framework of the general class of payo utility functions that both i are easy for the individual processes to try to maximize and ii have the property that if good values of them are achieved then we are assured of a high value of world utility These are the Aristocrat Utility and a new variant of the Wonderful Life Utility that was introduced in the previous COIN work We demonstrate experimentally that using these new utility functions can result in signi cantly improved performance over that of previously investigated COIN payo utilities over and above those previous utilities superiority to the conventional team game utility These results also illustrate the substantial superiority of these payo functions to the perhaps the most natural version of the economics technique of endo genizing externalities

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Evolving Optimal Populations with XCS Classi er Systems

This work investigates some uses of self-monitoring in classi er systems (CS) using Wilson's recent XCS system as a framework. XCS is a signi cant advance in classi er systems technology which shifts the basis of tness evaluation for the Genetic Algorithm (GA) from the strength of payo prediction to the accuracy of payo prediction. Initial work consisted of implementing an XCS system in Pop11 a...

متن کامل

Extensional and intensional collectives and the de re/de dicto distinction

Expressions designating collectives, such as “the committee” or “the ships in the port”, may be interpreted de re or de dicto, depending on context, according as they pick out collectives defined by their members or collectives defined by some criterion for membership. We call these E(xtensional)-collectives and I(ntensional)-collectives respectively, and in this paper we explore in depth the r...

متن کامل

Optimal Payoff Functions for Members of Collectives

We consider the problem of designing (perhaps massively distributed) collectives of computational processes to maximize a provided “world” utility function. We consider this problem when the behavior of each process in the collective can be cast as striving to maximize its own payoff utility function. For such cases the central design issue is how to initialize/update those payoff utility funct...

متن کامل

Project Contracting Strategies for Managing Team Dynamics

In this chapter we study a team dynamic problem in which a group of agents collaborate over time to complete a project. The project progresses at a rate that depends on the agents’ e↵orts, and it generates a payo↵ upon completion. First, we show that agents work harder the closer the project is to completion, and members of a larger team work harder than members of a smaller team both individua...

متن کامل

RELIABILITY-BASED MULTI-OBJECTIVE OPTIMAL DESIGN OF SPATIAL TRUSSES USING GAME THEORY AND GA

This paper introduces a reliability-based multi-objective design method for spatial truss structures. A multi-objective optimization problem has been defined considering three conflicting objective functions including truss weight, nodal deflection, and failure probability of the entire truss structure with design variables of cross sectional area of the truss members. The failure probability o...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2001