Consensus Theory for Cognitive Agents’ Unstructured Knowledge Conflicts Resolving in Management Information Systems

Abstract

Management information systems of distributed nature, play a vital role in any kind of business organizations’ activity. The multi-agent systems, based on cognitive agent architecture, deserve special attention in this class of systems. They allow not only to access to the information and quick search for interesting us information, its analysis and drawing conclusions, but also, in addition to responding to stimuli from the environment, have the cognitive ability to learning through empirical experience gained through direct interaction with the environment. It, in turn, allows for the automatic generation of variants of decisions and, in many cases, even taking and putting into action the decisions. The big problem currently, however, turns out to be the processing of unstructured knowledge in systems of this kind. In contemporary companies, unstructured knowledge is essential, mainly due to the possibility to obtain better flexibility and competitiveness of the organization. Therefore, unstructured knowledge supports structured knowledge to a high degree. Simultaneously, one must note that the most prevailing phenomenon is a conflict in unstructured knowledge. It is extremely difficult to resolve conflicts of this kind properly. However, it is also very important, since it can improve the operation of management information system and, consequently, help the organization that employs the system become more flexible and competitive. The main aim of this work is to develop a formal method to resolve conflicts in unstructured knowledge of cognitive agents in management information systems employing the consensus theory. The first part of this work presents an analysis problems related to management information systems and unstructured knowledge processing in these systems. Next, the cognitive agents are characterized with particular emphasis on unstructured knowledge processing. The use of consensus theory in unstructured knowledge conflicts resolving have been characterized in the third part of the work. The last part presents the developed method for cognitive agents’ knowledge conflicts resolving. The correctness of the method was verified using the prototypes of the agents helping to invest in the Forex market and processing user opinions about products and services.