基于离散粒子群优化的MPSoC节能调度算法

为了降低多核片上系统MPSoC在应用中的能耗,在MPSoC上提出了基于优化离散粒子群算法的节能任务调度算法.通过比例选择算子生成初始种群,以任务在MPSoC上不同内核执行的能耗作为解空间,粒子群在整个解空间上搜索最低能耗调度方案,并在算法中优化了粒子群算法的局部早熟问题,使算法性能进一步提升.仿真实验表明:基于优化离散粒子群算法的节能调度算法与常用的3种调度算法相比,能耗得到了降低,且算法的截止期错失率并没有升高,保证了算法的整体性能.

MPSoC Energy Saving Scheduling Algorithm Based on Discrete Particle Swarm Optimization

In order to reduce the energy consumption of the MPSoC in the application,a energy saving task scheduling algorithm based on the optimization of discrete particle swarm optimization is proposed in MPSoC.The algorithm generates the initial population by the proportional selection operator,and takes the energy consumption of different cores on the MPSoC as the solution space,search for the lowest energy consumption scheduling scheme in the whole solution space and optimize local premature problem of the particle swarm optimization algorithm.Through these improve the performance of the algorithm.The simulation experiments show that the energy of saving scheduling algorithm based on the optimized discrete particle swarm optimization algorithm is lower than that of the common three algorithms,and the deadline miss rate of the algorithm does not increase,that ensure the overall performance of the algorithm.