Evolutionary Scientific Workflows

Evolutionary computing has become an alternative to classical computational models since it can solve, efficiently, hard scientific problems in polynomial time. Workflows support the automation of scientific processes, providing mechanisms that underpin modern computational science. However, the traditional workflows offer little support to evolutionary computing. Thus, to bridge this gap, this paper describes the VisPyGMO. It extends the existing scientific workflow management systems by offering a range of reusable evolutionary algorithms modules. Besides, we present a use case using evolutionary algorithms in VisTrails system to analyze more than 20 years of historical data retrieved from Kaggle platform and to predict pesticide consumption in fresh foods. The results presented demonstrate the feasibility of the proposal and demonstrate that significant optimization can be achieved with an evolutionary approach for the optimization of state-of-the-art scientific workflows.