---
title: "A Dynamic Theory of Incentives in Multi-Agent Systems"
authors:
  - "Yoav Shoham"
  - "Katsumi Tanaka"
doi: ""
url: "https://ijcai.org/proceedings/1997/92"
pdf: "https://ijcai.org/proceedings/1997/0092.pdf"
venue: "Proceedings of the Fifteenth International Joint Conference on Artificial Intelligence"
year: 1997
pages: "626-630"
track: "Distributed Artificial Intelligence"
keywords:
---

# A Dynamic Theory of Incentives in Multi-Agent Systems

Yoav Shoham, Katsumi Tanaka

*Proceedings of the Fifteenth International Joint Conference on Artificial Intelligence*, pages 626-630.

## Abstract

Motivated by problems that plague the acceptance of groupware products, as well as the challenge of designing effective multi-agent systems, we investigate the role of incentives in "public goods'' settings, that is, settings in which the value of the system is supplied by its members. In such settings one encounters a prisonersdilemma-like problem, in which each individual has a disincentive to contribute, but much benefit from the contributions of others. The natural result is that, unless special care is taken to engineer correct incentives, no contribution is made by anyone and the system has no value. Drawing on the literature on mechanism design and public goods in economics, we present a model in which to reason about such settings. Our model has several novel features, the most important of which is its dynamic component; the model describes how the value of the system is determined over time as a result of its use. We provide several results about the model, including the identification of a critical mass point, which is a value of the system that, if reached, is guaranteed to henceforth enable the system to engineer incentives that lead to contribution to (and thus added value of) the system.