Evaluating Degree of Dependency from Domain Knowledge Using Fuzzy Inference System

Abstract

Inter-agent communication is one of the main concerns of agent oriented requirements engineering that is delineated as managing inter-dependencies and interaction among various agents performing collaborative activities. To carry out cooperative activities, tasks are disseminated and delegated to other agents with a rationale of sharing mutual expertise and potential. So an agent may be dependent on other agent for accomplishing a goal or for want of a resource to achieve that goal. This requires an agent to quantify the dependency needs termed as Degree of Dependency (DoD) to realize whether it should delegate the task to another agent, if yes, then to whom so that overall quality of Multi-Agent System (MAS) is not compromised. To evaluate DoD, this work employs domain knowledge that is an inclusive knowledge of an environment containing the business rules, credentials and reports to understand the business needs precisely. As the domain knowledge may be fuzzy and uncertain, Fuzzy Inference system is utilized to evaluate DoD from domain knowledge. This will assist the developer to address inter-agent coordination issues without squandering resources and hence in building MAS of high quality.