Multiple Ant Colonies Optimization for Load Balancing in Distributed Systems

Ant colony optimization (ACO) has proved its success as a meta-heuristic optimization in several network applications such as routing and load balancing. In this paper, a proposed ACO algorithm for load balancing in distributed systems will be presented. This algorithm is fully distributed in which information is dynamically updated at each ant movement. Multiple colonies paradigm will be adopted such that each node will send a colored co lony throughout the network. Using colored ant colony helps in preventing ants of the same nest from following the same route, and hence enforcing them to be distributed all over the nodes in the network. Each ant acts like a mobile agent that carries newly updated load balancing information to the next visited node. Finally, the proposed algorithm will be compared with the standard work-stealing algorithm.