Parallel cooperating A-Teams solving instances of the Euclidean planar travelling salesman problem

Abstract

The paper proposes an architecture for solving difficult combinatorial optimization problems through cooperation of several A-Teams working in parallel. The idea extends JABAT environment through integrating the team of asynchronous agent paradigm with the island-based genetic algorithm concept. The proposed architecture has been tested on instances of the Euclidean planar travelling salesman problem. Computational experiment has confirmed that the proposed architecture is competitive in terms of the quality of solutions in comparison with using traditional, non-cooperating A-Teams. © 2011 Springer-Verlag.