ANALYZING AND MODELING SOFTWARE AGENTS
Fan Xiao · 2000
Agents in AI are entities with the features of reactiveness and, pro activeness,which are embedded in and perceive changing, uncertain worlds, and act in it. In recent years, substantial progress has been made both in theoretical and in practical aspects of agents and multi agent systems. As yet, however, there are not many examples of the successful application of agent system technologies on a significant scale. One of the main obstacles, we think, is due to the absence of methodologies that allow system complexity to be effectively managed. If multi agent systems are to become widely accepted as a basis for large scale commercial and industrial applications, it is essential to research adequate design methodologies and modeling techniques. Only by this can we guarantee that such systems are sufficiently reliable and maintainable, and allow their design, implementation, and maintenance to be carried out by software analysts and engineers rather than AI researchers. By far, various Object oriented approaches may be better solutions amongst the methodologies that have been developed for the design, specification, and programming of conventional software systems. However, agents are usually significantly more complex than typical objects both in their internal structure and in the behaviors they exhibit. We cannot directly apply OO methodologies to agent systems. Since the existing formalisms for describing and reasoning about agents do not provide adequate support for the process of agent design, we try to develop suitable Agent oriented methodologies. By building upon existing, well understood techniques, we take advantage of the OO approach and aim to develop a methodology that will be easily learnt and understood by those who are familiar with OO paradigm. Based on information agents in open distributed environments, this paper offers an agent oriented analysis—ARC, illustrates the idea of three level modelling: object based components, active objects and agents, and tries to provide a feasible analysis and modeling method for Agent based systems.