Agents for the masses

Abstract

Is it possible to make development of sophisticated agents simple enough to be practical? This article discusses the authors' theoretical and tool-creation efforts to answer that question in the affirmative. Building sophisticated agent-based systems of the future will require research advances on at least three fronts: we must continue work on agent theory so that many currently unanswered questions about the scope and limitations of alternative approaches to agent design can be addressed; we must make agent frameworks and infrastructure powerful, interoperable, and secure enough to support robust large-scale coordinated problem-solving activity; and we must develop new sorts of tools to help non-specialists unlock the power of agent technology.