基于量子蚁群算法的片上网络映射研究*

随着片上网络的兴起和发展,针对带宽和时延约束下实现低功耗成为其设计的焦点之一。为此,提出一种基于量子蚁群映射算法的方法来解决片上网络设计中使IP核映射的通信功耗最小化问题。该算法改变蚁群算法中信息素的释放方式,采用量子优化算法中的量子概率幅代替,信息素的更新则通过使用量子相位旋转的方式,实现蚂蚁信息素的自适应更新,用以有效的降低蚁群算法容易早熟收敛的情况。通过实验对比研究,该算法在快速搜索和全局寻优能力上,均优于蚁群算法。

Research on NoC mapping based on quantum ant colony algorithm

With the developing of Networks on Chip, according to the bandwidth and delay constraints to achieve low power consumption has become one of the hotspots of the design. This paper presented a Quantum Ant Colony Algorithm (QACA) strategy to map applications on Networks on Chip. This QACA used quantum bit in quantum evolutionary algorithm to replace the ant colony pheromone. Based on the adaptive phase rotation strategy, made the pheromone update dynamically to reduce the premature convergence of the ant colony algorithm effectively. Experimental results show that the proposed algorithm is better than the ant colony algorithm both at the ability of search capability and global optimization.