关键词: 片上网络（NoC）, 映射, 云模型, 遗传算法

Abstract: NoC mapping is an important step of the NoC design. The quality of the NoC mapping determines the effect of QoS constraints and communication power consumption. This paper proposes a NoC mapping solutions based on cloud adaptive genetic algorithm. The algorithm is based on cloud model to improve the traditional genetic algorithm, taking a new method to improve crossover probability and mutation probability. Combined with the practical problems of NoC mapping, in constraints of power consumption and delay conditions, the power consumption of the mathematical model under the delay restriction of the NoC mapping is established. The experiment results show that the method has obtained favorable effect and reduced the power consumption of communication.

Key words: Network-on-Chip（NoC）, mapping, cloud model, genetic algorithm