Methodological Bases of Decision Support for Resource Management in Complex Systems

Abstract

This paper discusses the methodological basis of decision support for resource managing in complex systems. Existing approaches to resource management use mostly rigid mathematical tools and data of a numerical nature. In this case, ill formalized information about the features of the subject domain is often not taken into account. The proposed methodology is a sequence of implementation of certain stages, based on appropriate approaches, methods and principles. Its distinctive feature is the integration of multi-agent technologies and ontological models, which allows to take into account ill formalized information about the subject domain. A formal model of a multi-agent system for resource managing in complex systems is proposed. It allows modeling the resource allocation process and analyzing the results of such modeling. The model includes two main types of agents: resource agents and resource consuming agents, as well as their corresponding knowledge bases. This model allows solving resource allocation problems taking into account semantic constraints and dynamically changing external environment. A scheme of a resource management system in complex systems is proposed. It allows for constructing resource allocation plans and monitoring their execution. The results of this monitoring are used further to improve the efficiency of resource management.