Intelligent scheduler with fuzzy techniques and evolutionary management in heterogeneous computing systems
Phạm Hồng Hạnh · 2002
The paper describes an approach to solving the job scheduling problem in heterogeneous computing systems. In the proposed model, we use a fuzzy function to estimate and code the system's states. Evolutionary management is applied to find the optimum schedule using offline knowledge or online algorithms. The new scheduling system provides significant improvement in scheduling quality and time, and is capable of self learning and self adjustment.