一种基于神经网络的多层通孔最小化方法  

A Neural Network Approach to Multi Layer Constrained Via Minimization

作　　者：马琪[1] 严晓浪[1] 胡卫明[1] 

机构地区：[1]杭州电子工业学院ICCAD所

出　　处：《微电子学》1997年第1期21-25,共5页Microelectronics

基　　金：电子科学研究院预研项目

摘　　要：在多层布线的线段－相交图模型基础上，利用Ｈｏｐｆｉｅｌｄ人工神经网络理论，通过把通孔数目这个优化目标与Ｈｏｐｆｉｅｌｄ网络能量函数相联系的方法来解决多层布线通孔最小化问题。算法考虑了许多来自实际的约束。结果证明，这是一个有效的算法。A new approach to multi layer constrained via minimization is presented.The new approach based on the segment crossing graph model and Hopfield neural network theory.The minimization is achieved by correlating the number of vias with the energy function of Hopfield networks.In addition,many physical constraints are taken into consideration.It has been shown that this algorithm is very efficient and has the advantage of fast converging.

关 键 词：VLSI CAD 通孔最小化 多层布线 

分 类 号：TN470.2[电子电信—微电子学与固体电子学]