The right agent (architecture) to do the right thing

Academic and industrial system designers who consider using agent technology to solve an application problem are faced with a wide variety of agent paradigms: There are deliberative agents, reactive agents, interacting agents, hy brid agents, layered agents, believable agents, mobile agents, software agents, softbots - the list could well be prolonged. Also, within each paradigm, the user can select between different architectures and systems, making the actual choice a complex and difficult endeavor. The objective of this paper is to assist readers in deciding which agent architecture to choose for a specific application. We approach this objective in three steps. First, we identify application areas for agent technology starting from the examples presented in the first part of this paper. Then, based on the characteristics of different classes of applications identified in the first step, we propose a classification of agents according to different classes of applications. Based on this classification, the third step is to provide rules of thumb to help a software engineer or system designer decide which agent architecture (or which class thereof) is likely to be appropriate for a certain class of applications.