Learning to Participate through Trading of Reward Shares

TL;DR

This paper introduces a stock market-inspired mechanism where autonomous agents trade reward shares to promote cooperation, leading to role development and better social dilemma resolution in multi-agent systems.

Contribution

It proposes a novel reward sharing method based on trading shares, enhancing cooperation and role formation among agents in complex environments.

Findings

Promotes cooperative policies in social dilemmas

Enables role development and subtask division

Effective in extended domains

Abstract

Enabling autonomous agents to act cooperatively is an important step to integrate artificial intelligence in our daily lives. While some methods seek to stimulate cooperation by letting agents give rewards to others, in this paper we propose a method inspired by the stock market, where agents have the opportunity to participate in other agents' returns by acquiring reward shares. Intuitively, an agent may learn to act according to the common interest when being directly affected by the other agents' rewards. The empirical results of the tested general-sum Markov games show that this mechanism promotes cooperative policies among independently trained agents in social dilemma situations. Moreover, as demonstrated in a temporally and spatially extended domain, participation can lead to the development of roles and the division of subtasks between the agents.