The Agent Oriented Multi Flow Graphs Specification Model
I. D. · InTech eBooks · 2011
The term agent is widely used in many areas with diverse meaning. In this work, agents have been used as they have been defined in the context of AI. In this context, many researchers use notions to describe an agent that are normally applied to humans, such as knowledge, beliefs, desires, intentions, capabilities, choices, commitments or obligation, (Shoham, 1993; Cohen & Levesque, 1990; Rao & Georgeff, 1991a) or even emotions (Bates, 1994). Wooldridge in (Wooldridge, 1999) uses a broader definition and argues that an agent denotes a hardware or software-based system that exhibits properties such as autonomy, social ability, reactivity, and pro-activeness. Several other agent properties that have been suggested include mobility, veracity, benevolence and rationality. This work shares the need of the above properties and in order to express them an agent is considered as an intentional system based on the description of the philosopher Daniel Dennett. He proposed the intentional stance from which systems are ascribed mental qualities. Using this term, he describes entities “whose behaviour can be predicted by the method of attributing belief, desires and rational acumen” (Dennett, 1987). The intentional stance deals mainly with the behaviour of the described entities and helps to avoid paradoxes and clashes of intuition (Russell & Norvig, 2003). Its main disadvantage is that it cannot distinguish among implementations since any given behaviour can be implemented in many different ways and does not imply anything about the system’s internal workings or representation. However, from an engineering point of view, it allows the better understanding and consequently the description of a complex system and for this reason is adopted by this work. Concluding that an agent is a system that is most conveniently described by the intentional stance, Wooldridge and Jennings (Wooldridge & Jennings, 1995) consider that the appropriate attitudes for representing agents are classified in information attitudes (belief and knowledge) which are related to the information that an agent has about the world it occupies, and pro-attitudes (desire, intention, obligation, commitment, choice etc.), which in some way guide the agents’ actions. More than one agent, who cooperate or deliberately act together in order to accomplish a task (usually common) constitute a multi agent system. Multi agent systems are inherently complex since “concurrency, problem domain uncertainty and non-determinism in