摘要： 布局是VLSI布图设计中的关键环节，通常采用随机优化算法。该文采用遗传算法(GA)与模拟退火法(SA)相结合的搜索算法实现VLSI门阵列模式布局，利用遗传算法进行全局搜索，模拟退火法进行局部搜索。进化过程中采用精英保留策略，并对进化结果进行有选择的模拟退火操作，这样既加强了局部搜索能力又防止陷入局部最优。在复合布局目标函数中引入对最长线网的惩罚，其收敛速度比以总线长度为单一目标函数的要快。在交叉操作中，对交叉位置的选择采用了一种新的策略，增加了交叉的有效性。实验表明，此算法与简单遗传算法相比，有效地提高了全局搜索能力。

关键词: VLSI布局, 遗传算法, 模拟退火法

Abstract: A new algorithm for regular ICs’ placement problem is presented. It is resulted from the combination of the genetic algorithm and simulated annealing algorithm, the genetic algorithm is served as the main flow of the new algorithm for global search, simulated annealing algorithm adjusts the optimization population to avoid trapping in the local optimum. In the genetic algorithm, a new crossover operator involved Gauss probability to determine the crossover slot is described. The new algorithm has been run on a variety of test instances; the result is superior to that of simple GA.

Key words: VLSI placement, Genetic algorithm, Simulated annealing