Agent system development method based on agent patterns

Abstract

As wide-area open networks such as the Internet and intranets grow larger, agent technology is attracting more attention. Agents are units of software that can deal with environmental changes and the various requirements of open networks through features such as autonomy, mobility, intelligence, cooperation, and reactivity. However, since the usual development method of the agent systems is not sufficiently investigated, the technology is not yet widespread. This paper proposes a method of agent system development based on agent patterns that represent typical and recurring structures and behaviors of agents. The agent patterns are classified according to their appropriate architectural levels and the degree to which they depend on specific agent platforms. Our method enables developers to design agent systems efficiently since they can construct complicated system architectures and behaviors by dividing the design process into two architectural levels and applying the appropriate agent patterns. In addition, the higher level designs are independent of specific agent platforms and can be therefore easily reused.