Skip to Main Content
Since the nonlinearity of the battery behavior and its dependence on the characteristics of the discharge profile, maximizing battery lifetime is particularly difficult problem for mobile computing devices. Dynamic voltage scaling (DVS) is a promising technique for battery-powered systems to conserve energy consumption. Even if information about task periodicity or a priori knowledge about the task set is known, DVS scheduling problem where the target processor operates at discrete voltage is well known to be NP-hard in general. In this paper, efficient scheduling algorithms for both aperiodic and periodic task sets on DVS systems are presented. The proposed heuristics algorithms based on GA using a charge-based cost function derived from the battery characteristics. The efficiency of the proposed algorithm has been verified by shown superior results on synthetic examples of periodic and aperiodic tasks which were excerpted from comparative work or were generated randomly, on uniprocessor or multiprocessor platforms. Our experimental results demonstrating that the proposed scheduling algorithm significantly reduces up to 19% of dynamic energy consumption compared with a past approach.