Energy-Aware Scheduling of Independent Tasks in Computational Grids

Abstract

This chapter introduces the application of the Hierarchical Genetic Strategy-based Grid scheduler (HGS-Sched) to the energy-aware independent batch scheduling problem in Computational Grids (CGs). The Dynamic Voltage Scaling (DVS) methodology is used for both scaling the power supply of the grid resources and reducing the cumulative power energy utilized by the grid computing machines. Two implementations of HGS-Sched—with elitist and struggle replacement mechanisms respectively—are defined and empirically evaluated. The effectiveness of the hierarchical schedulers are compared with the quality of single-population Genetic Algorithms (GAs) and Island GA models for four CG significant scenarios in static and dynamic modes. The simulation results show that meta-heuristic grid schedulers can significantly reduce the energy consumption in the system as well as be easily adapted to various scheduling scenarios.