许川佩, 侯红菊. 基于云自适应粒子群算法的NoC路径分配研究[J]. 微电子学与计算机, 2011, 28(9): 13-16,21.
引用本文: 许川佩, 侯红菊. 基于云自适应粒子群算法的NoC路径分配研究[J]. 微电子学与计算机, 2011, 28(9): 13-16,21.
XU Chuan-pei, HOU Hong-ju. Research of NoC Path Allocation Based on Adaptive PSO Based on Cloud Theory[J]. Microelectronics & Computer, 2011, 28(9): 13-16,21.
Citation: XU Chuan-pei, HOU Hong-ju. Research of NoC Path Allocation Based on Adaptive PSO Based on Cloud Theory[J]. Microelectronics & Computer, 2011, 28(9): 13-16,21.

基于云自适应粒子群算法的NoC路径分配研究

Research of NoC Path Allocation Based on Adaptive PSO Based on Cloud Theory

  • 摘要: 在云自适应粒子群算法基础上提出一种新的云自适应更新函数并将其应用在2DMesh拓扑的片上网络中, 来优化静态通讯分配结果.该更新函数的参数随着粒子适应度值和拓扑规模的不同而变化, 并注重全局搜索和局部收敛的结合作用, 使两者达到最佳结合点, 避免了单一的全局搜索和局部收敛的弊端.实验证明, 这种方法在拓扑规模越大时, 效果越明显, 与改进的粒子群算法相比较结果优达25.74%.

     

    Abstract: In this paper, based on adaptive Particle Swarm Optimization (PSO) based on cloud theory, a new adaptive update function based on cloud theory is proposed and applied in the 2D-Mesh topology to optimize distribution of static communication.The parameters of the function varies with the fitness value of particle and the size of topology.The function focus on global and local convergence of binding, so that which achieve the best combination of point to avoid disadvantages of single global search and single local convergence.Experiments show that the effect is more obvious when the size of topology is larger and excellent results is up to 25.74%when comapared with PSO.

     

/

返回文章
返回