Comparing Market and Token Based Coordination

Abstract

Many coordination algorithms claim to be general, implying that they can be used to coordinate agents in a variety of domains. However, little work has been done to quantitatively compare distinctly different approaches to coordination across a range of domains. In this paper, we present a detailed comparison of two published coordination algorithms, performed in an abstract coordination simulation environment that allows extensive, quantitative experimentation. The simulator is to used to compare two distinct approaches to coordination, token-based coordination and market based coordination. The results largely show the generality of different approaches, but performance and performance tradeoffs varies greatly across domains.