结合模拟退火和遗传算法各自的特点，提出了一种混合算法，用以设计阵列型惯性约束核聚变（ＩＣＦ）均匀照明系统中二元阵列单元的位相分布。混合算法充分利用了遗传算法的并行性及保留一定历史信息的特性，并用模拟退火的温度参数控制收敛性。模拟运算表明，混合算法具有较高的效率及寻优可靠性。用本算法设计的均匀照明阵列，可以得到顶部均匀性为３．２％（ｒｍｓ）、能量利用率接近９０％的焦斑 Combining the respective characteristics of simulated annealing and genetic algorithms, a hybrid algorithm is proposed to design the phase distribution of binary array elements in an arrayed inertial confinement fusion (ICF) homogeneous illumination system. The hybrid algorithm makes full use of the parallelism of genetic algorithm and preserves the characteristics of a certain historical information, and uses the temperature parameter of simulated annealing to control the convergence. The simulation results show that the hybrid algorithm has higher efficiency and better reliability. With the uniform illumination array designed by this algorithm, it is possible to obtain a focal spot with a top uniformity of 3.2% (rms) and an energy efficiency of nearly 90%