An energy-efficient optimal algorithm based on static voltage scheduling for battery-powered DVS systems

Abstract

Energy saving is of great importance and energy-efficient optimal is considered as a critical element in battery design. In this paper, a novel static voltage scheduling algorithm for battery-aware dynamic voltage scaling (DVS) systems is presented. The proposed algorithm pays fully attention to the relationship of the factors (the increase of the slack time and discharge) related to battery consumption by taking the nonlinearity characteristics of the battery into account. We focus on the decreasing speed of the cost function which is a critical parameter in energy saving. The proposed algorithm can get the fastest speed in cost function decreasing and finally return a profile in accordance with the profile obtained by brute-force searching. Aperiodic and periodic tasks on uniprocessor and multiprocessor are considered in the proposed algorithm to make our algorithm widely used.