Architectural Design of a GAIA Based Multi Agent Enterprise Knowledge Management System

Abstract

An agent is a computer system that is situated in some environment, and that is capable of autonomous action in this environment in order to meet its design objectives. An agent is social, reactive and proactive. Multi agents systems (MAS) are used to solve certain problems which are difficult or impossible for stand-alone systems like financial portfolio newlinemanagement, supply chain management etc. newline newlineFor the implementation of MAS, it s necessary to use any Agent Oriented Software Design methodology which can help in analyzing and designing the system. The research work uses Gaia, extended for closed MAS for designing the MAS. newline newlineKnowledge is the utilization of information and data, along with the people s skills, competencies, ideas, intuitions and motivations. An Enterprise Knowledge management system (EKMS) attempts to build a range of strategies to facilitate the definition, identification, capture, organization and dissemination of knowledge across the organizational community. EKMS have different phases as Knowledge acquisition, Knowledge creation, Knowledge codification, Knowledge storage, Knowledge retrieval,Knowledge sharing and Knowledge use. newline newlineAn EKMS is best implemented by a network of agents. An efficient Multi Agent Enterprise Knowledge Management System (MAEKMS) is designed using fuzzy c-means clustering,case based reasoning and knowledge beads. The proposed architectural design of MAEKMS is then applied to various diverse domains of IT industry, health care services and ecommerce. newlineThe MAEKMS for e-commerce is dealt in detail which includes its analysis and evaluation using linear, polynomial and multiple regression techniques.