论文部分内容阅读
应用前馈神经网络解深亚微米 MOSFETs反型层量子效应模型 .此方法需要用神经网络解微分方程的特征值问题以及积分方程 .首先以此方法解具有 Morse势函数的 Schro¨ dinger方程来验证此方法的精度 ,然后 ,以此方法解 Schro¨ dinger方程和 Poisson方程的近似模型——三角势阱模型 .该方法易于实现 ,便于扩展到二维模型 ,方便模拟 .
Applying the feedforward neural network to solve the inverse layer quantum effect model of deep submicron MOSFETs, this method needs neural network to solve the eigenvalue problem and integral equation of differential equation.Firstly, Schro¨dinger equation with Morse potential function is verified by this method The accuracy of this method is then used to approximate the Schroßinger equation and the approximation model of the Poisson equation - the triangular well model. This method is easy to implement and easy to extend to 2D models for simulation.