Abstract: To solve the hardware/software partitioning problem in embedded system，this paper proposes a New Genetic Simulated Annealing algorithm（NGSA） which based on analysis of genetic algorithms and simulates annealing algorithm the main advantages and disadvantages.The genetic algorithm integrates the simulated annealing idea；niche technology is introduced to maintain population diversity；and the Metropolis criterion with the formation of new groups to improve the quality of group.Experimental results show that the algorithm has strong climbing ability and global search capability，and the fitness value is significantly improved than genetic algorithm and simulated annealing algorithm.

Key words: embedded system, hardware/software partitioning, genetic algorithm, simulated annealing algorithm, niche technology

摘要： 针对嵌入式系统软硬件划分问题，在分析遗传算法和模拟退火算法的主要优缺点的基础上，提出了一种新的小生境技术改进的遗传模拟退火算法（NGSA），在遗传算法中融入模拟退火思想，同时引入小生境技术，保持群体的多样性；并采用Metropolis 法则形成新群体，改善群体的质量。实验结果证明该算法具有很强的爬山能力和全局搜索能力，与遗传算法（GA）和模拟退火算法（SA）相比适应度明显提高。

关键词: 嵌入式系统, 软硬件划分, 遗传算法, 模拟退火算法, 小生境技术