Metrics for multiagent systems

Abstract

A Multiagent System (MAS) is a software paradigm for building large scale intelligent distributed systems. Increasingly these systems are being deployed on handheld computing devices that rely on non-traditional communications mediums such as mobile ad hoc networks and satellite links. These systems present new challenges for computer scientists in describing system performance and analyzing competing systems. This chapter surveys existing metrics that can be used to describe MASs and related components. A framework for analyzing MASs is provided and an example of how this framework might be employed is given for the domain of distributed constraint reasoning. © 2009 Springer-Verlag US.