通孔最小化的一种神经网络方法

A Neural Network Approach to Constrained Via Minimization

本文在布线的群图模型基础上，利用离散型Hopfield神经网络解决群图的最大割问题，并着重论述了如何跳出局部最优点的问题，从而较好地解决了双层布线通孔最小化问题。算法考虑了许多来自实际的约束，并进行大量的布线实例验证。

In this paper, we present a new approach for twolayer constrained via minimization by means of discretehopfield neural network on the basis of weighted cluster graph model. In addition, many physical constraints are taken into consideration here. Accordingto the results, our algorithm is verified to be very efficient and encouraging.