Agent-based Management and Optimization System for Distributed Computing

C. Chira and D. Dumitrescu (Romania)


Distributed Computing, Multi-Agent Systems, Natural Computing, Metaheuristics, Stigmergy


Distributed computing systems aim to connect human and information resources in a transparent, open and scalable way. Users dispersed over a computer network have to virtually collaborate using distributed tools. Multi-agent systems and ant colony optimization are investigated with the aim of compiling a successful approach to solve distributed computing problems. Composed of several interacting agents, multi-agent systems have the ability to provide solutions to efficiently manage domains in which information resources and expertise are inherently distributed. An agent-based architecture is proposed to support distributed computing by enabling interoperation among distributed resources, facilitating knowledge integration and addressing inherent complex problems. The multi-agent approach addresses load balancing and resource management problems. A novel aspect of the proposed architecture refers to the employment – within a a particular agent subsystem – of nature-inspired metaheuristics (such as ant colony optimization techniques) to solve various complex problems inherent to a distributed computing environment.

Important Links:

Go Back