Abstract
This paper presents a method for agents' knowledge representation by using semantic network with node and links activation level defined on the instance, concept, relation and axiom level. The first part shortly presents the state-of-the-art in the considered field; next, the CIMIS prototype is shortly characterized; the formal definition of a method for agents' knowledge representation is presented in the last part of paper. I. INTRODUCTION ONTEMPORARY the entire economy is based on information and knowledge, therefore companies must employ systems which support the knowledge management process taking into consideration the risk and uncertainty of economic decisions. Often the integrated management information systems (IMIS) are used in this purpose. They are characterized by full integration both at the system/application level and the business process level. Note, however, that the properties of contemporary IMIS are becoming more and more inadequate. Apart from collecting and analyzing data and generating knowledge, the system should also be able to understand the meaning of phenomena occurring around the organization. It is becoming more and more necessary to make decisions based not only on knowledge but also on experience, thus far regarded as purely human domain (4). In order to accomplish tasks set by IMIS, a multi-agent system can be used consist of several cognitive agents. Not only do they enable quick access to information and quick search for the required information, its analysis and conclusions, but also, besides being responsive to environment stimuli, they have cognitive abilities that allow them to learn from empiric experience gained through immediate interaction with their environments (15), which consequently allows a number of decision versions to be automatically generated and to make and execute decisions. As a result of its running, the agent obtains knowledge of the environment in which it operates. If we desire to use this knowledge then it must be represented in the form of a
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have