An enabling framework for parallel optimization on the computational grid

In this paper, we present ParadisEO-CMW, an extension of the open source ParadisEO framework, originally intended to the design and deployment of parallel hybrid meta heuristics on dedicated clusters of SMPs. Coupled with the Condor-MW library, it enables the execution of such parallel applications on volatile heterogeneous computational resources. The motivations, architecture and main features will be discussed. The framework has been tested by tackling a real-world NP-hard problem: feature selection in near-infrared spectroscopic data mining. It has been resolved by deploying a multi-level parallel model of evolutionary algorithms. Experimentations have been carried out on more than one hundred PCs originally intended for education. The obtained results are convincing, both in terms of flexibility and easiness at implementation, and in terms of efficiency and quality of provided solutions at execution.