利用三状态Hopfield模型优化神经网络双极二值化互连

Optimization of Binary and Bipolar Interconnection by Using the Hopfield Model with Tristates Neurons

针对光互连精度较低的弱点，提出了以正交化算法为目标函数的三值（±１，０）互连权重的神经网络优化方法。计算机模拟和光学实验结果表明该方法能显著地提高三值互连神经网络模型的性能。

In order to overcome the weakness of poor accuracy of optical interconnection, an optimization method of binary and bipolar associative model is proposed by using a neural network and a target function based on orthogonal projection algorithm. Computer simulation and experimental results show that the performance of the optimum model is greatly improved.