تقسیم وظایف آگاه از انرژی روی پلتفرم های چند پردازنده ناهمگن

Efficient task partitioning plays a crucial role in achieving high performance at multiprocessor platforms. This paper addresses the problem of energy-aware static partitioning of periodic realtime tasks on heterogeneous multiprocessor platforms. A Particle Swarm Optimization variant based on Min-min technique for task partitioning is proposed. The proposed approach aims to minimize the overall energy consumption, meanwhile avoid deadline violations. An energy-aware cost  function is proposed to be considered in the proposed approach. Extensive simulations and comparisons are conducted in order to validate the effectiveness of the proposed technique. The achieved results demonstrate that the proposed partitioning scheme significantly surpasses previous approaches in terms of both number of iterations and energy savings. Keywords: Task Partitioning, Task Assignment, Heterogeneous Multiprocessors, Particle Swarm Optimization, Min-min.