JADE-based distributed Ant Colony Optimization on a computer cluster

In this paper we present the initial experimental results that we obtained with deploying our distributed agent-based system for Ant Colony Optimization (ACODA) on a computer cluster. The novelty of ACODA consists in agent-based modeling and distribution of the problem environment that is explored by the ants to determine an optimal solution. The effect of this approach is that ants' migration is reduced to asynchronous messages exchanged between the agents that compose the problem environment. The deployment of ACODA on a computer cluster benefits from an automated setup that simplifies the installation and configuration of the necessary software packages.