Lifetime enhancement of embedded systems during task scheduling through a multi-objective particle swarm optimization approach

In this paper, a task scheduling and mapping method based on multi-objective particle swarm optimization (MOPSO) algorithm is presented to improve lifetime reliability of multiprocessors systems on chip. This method considers power consumption temperature and performance along with the lifetime reliability due to the antagonistic relations of these parameters. These antagonistic and dependent relations make the design space exploration and optimization processes complex. The proposed method is based on MOPSO algorithm and starts with an initial population of candidate solutions which are generated randomly and represents valid task scheduling and mapping solutions. By changing the scheduling and mapping parameters during the MOPSO algorithm, new solutions are produced and the design space is explored based on the objective of the target problem of this method. Several experiments on random and real-life benchmarks are performed to verify the effectiveness of our proposed method. The results demonstrate the capability of the proposed method in effective design space exploration and generating the Pareto front. Moreover, comparisons to the related research show 35%, 23%, 19% and 3% improvements in performance, lifetime reliability, temperature, and power consumption on average.