Lagrange优化神经网络的原理及实现

人工神经网络是指模拟人脑神经系统的结构和功能,运用大量的处理部件,由人工方式建立起来的网络系统.近几年来,人工神经网络的研究工作十分活跃,取得了很大的进展,研究开发出了几十种神经网络的模型,出现了多种新型神经网络.阐述了Lagrange优化神经网络的原理和简单的电路实现,它克服了传统的基于罚函数的神经网络的缺陷,直接对不等式约束进行处理,降低了网络规模和复杂度,是一种新型的优化神经网络,并通过计算机仿真对其可行性进行了验证.

Lagrange optimization neural network principle and implementation

The artificial neural network is the network system by the artificial way to set up,which refers to the structure and function of the nervous system of human brain of simulation,using a large number of treatment parts.In recent years,the research work of the artificial neural network is very active and make very great progress,the model is studied which develops dozens of kinds of neural net-works,many kinds structure of new-type neural networks have appeared.A kind of Lagrange principle of optimizing neural network and realization of simple circuit is sketched,it has overcome the traditional defect based on that the neural network which punish function thought exists,deal with inequality restraint directly,reduce network size and complexity,a kind of new optimization neural network,verify his feasibility through computer numerical value emulation.