Energy-efficient task scheduling algorithm for mobile terminal

Dynamic voltage scaling (DVS) is one of the most effective methods to saving energy consumption for battery-operated embedded systems. In this paper, we propose a genetic optimization based task scheduling algorithm that minimizes energy consumption for the case when the task deadline times and execution cycles are given. We construct a general mathematical scaling model of the processor that can be varied either continuously or discretely. Based on this model, a task model and energy consumption model are also provided. According to these models, energy minimization is equivalent to select appropriate scaling coefficients for each task while meet their deadline constrains. A hybrid genetic algorithm, which involves very low computation cost, is introduced to determine the scaling coefficients. The scheduling algorithm is not restricted to scaling types of the processor, inter-task DVS or intra-task DVS, periodic or aperiodic tasks. Experimental results show that the algorithm can reach near 56% energy saving.