Implementation Model for Agents with Layered Architecture in a Transactional Database Environment

One of the reasons for attending to agent technology is the ever-growing complexity of information systems and the increasing difficulty to foresee and plan for all potentially arising situations. However, the full use of the whole strength of the technology, as well as the research efforts, still remains restricted to a narrow area of Distributed Artificial Intelligence. Unfortunately, this means that some pressing issues of the practical applications remain outside the scope of agent research. One of those issues is robustness in unreliable error-prone or involuntarily interacting environments. On the other hand, database technology has for a long time learned to deal with robustness by developing a suitable technology, transactions. In this paper, an implementation model for agents with layered architecture in a transactional database environment is proposed. The model overcomes the non-orthogonality in implementing the various agent functions together with the limitations imposed on concurrent execution of agent actions. Two examples are provided to demonstrate our objective.