Semantic and multi-agent technologies in problems of resource management in grid O.M. Pakhnina, engineer, A.O. Matiushko, student National aerospace university named after N. Ye. Zhukovsky “KhAI"
Grid is an emerging technology for enabling resource sharing and coordinated problem solving in dynamic multi-institutional virtual organizations. In the Grid environment, the resources may belong to different institutions, have different usage policies and pose different requirements on acceptable requests. Resource discovery is process of locating relevant resources based on application requirements of a user.
With the development of grid increasing difficulties in the coordination and allocation of resources appears which are associated with the dynamics and uncertainties of resource protection and job flow, the need to consider the interests of users and resource owners, etc. The most promising direction of improving the efficiency of resource management in grid, taking into account the mentioned above features, is using of multi-agent and semantic technologies. Semantic technologies can improve various aspects of Grid systems: from discovery and selection of computing resources and data storage, scheduling to intelligent user interfaces. Multi-agent systems, taking into account the benefits of the services, can automate the complete cycle of real-time resource management, including the harmonization of interests, coordinated interaction, dynamic scheduling and adaptive re-planning tasks. Using semantic technologies in the form of ontologies in multi-agent systems also has its advantages: a common understanding of the subject area, the successful interaction between agents, the possibility of knowledge statement, regardless of system implementation, simplifying the programming of agent’s behavior. Agents must implement the logic of access to resources via grid-services for interaction between applications, users, or other agents that to support semantic interoperability. This paper suggests the multi-agent simulation model of resource management in grid, with the possibilities of economic self-organization and collective adaptation model elements with their own interests and decision¬making processes, which allows for decentralized planning with flexible management of resources and the load in the face of uncertainty and dynamic environment. The combined use of semantic and multi-agent technology opens up broad prospects for further improvement of the basic elements of grid systems, making the task of research, coordination and allocation of resources in grid systems, less time-consuming, and its solution is more efficient.