Abstract
Conference Title: 2019 IEEE International Conference on Electro Information Technology (EIT) Conference Start Date: 2019, May 20 Conference End Date: 2019, May 22 Conference Location: Brookings, SD, USA Power consumption in the present day embedded systems have become a complicated problem to tackle. With the consumers wanting more powerful devices with high performance demands, the issue of power management has been researched extensively to find better solutions. With the introduction of multiprocessor systems, the problem of power management has been alleviated to a certain degree. However, there is still much more work needed to be done to conserve energy. With this introduction, task migration has taken the spotlight as the multiprocessor architectures allow the tasks to be migrated as demanded to reduce power waste and hot spots of the systems. This paper presents a comprehensive framework that includes task allocation heuristic, a task migration that is based on genetic algorithm and an EDF scheduling scheme to properly schedule the migratable tasks. With our proposed techniques, we have shown a promising increase in the energy savings compared to other methods with task migration in the literature.