论文部分内容阅读
本文提出了一个深亚微米条件下的多层VLSI／MCM有约束分层层分配的遗传算法。该算法分为两步：首先进行超层（X－Y平面对）分配，使各线网满足Crosstalk约束、且超层数目最少；然后进行各超层的通孔最少化二分层。与目前的层分配算法相比，该遗传算法具有目标全面、全局优化能力强等特点，是一种可应用于深亚微米条件下的ICCAD的有效分层方法。
In this paper, we propose a genetic algorithm with constraint layering in multi-layer VLSI / MCM under deep sub-micron conditions. The algorithm is divided into two steps: firstly, super-layer (X-Y plane pair) distribution is made so that each line network satisfies the Crosstalk constraint with the minimum number of superlattices; and then the through-holes of each super-layer are minimized. Compared with the current layer allocation algorithm, the genetic algorithm has the characteristics of complete target and strong global optimization, and is an effective ICMA hierarchical method which can be applied to deep sub-micron conditions.