传统NSGAⅡ算法在解决NoC映射问题时容易出现局部最优和种群多样性的问题。为此,提出一个均衡优化时延模型和一种基于Boltzmann-NSGAⅡ的映射算法,时延模型从宏观链路负载和单个节点排队时延进行优化,映射算法引入Boltzmann机制、多点交叉算子、随机变异算子,以保持种群多样性并提高局部搜索能力。实验结果表明,Boltzmann-NSGAⅡ算法能获得更好的映射效果,最优解集的收敛性和均匀性更好。 The traditional NSGA Ⅱ algorithm is prone to local optimization and population diversity in solving NoC mapping problems. Therefore, a balanced optimization delay model and a mapping algorithm based on Boltzmann-NSGAⅡ are proposed. The delay model is optimized from the macro link load and the queuing delay of a single node. The mapping algorithm introduces the Boltzmann mechanism, the multi-point crossover operator, Random mutation operator to maintain population diversity and improve local search capabilities. Experimental results show that the Boltzmann-NSGAⅡ algorithm can achieve better mapping effect, and the convergence and uniformity of the optimal solution set are better.