chapter4

Source code for Hopfield character Recognition

1D CNN和2D CNN网络的仿真实现，包含完整的子函数分解，采用matlab编写(demo_cnn_auto- demo for Autonomous CNNs with a symmetric A template solve_cnn2d- solve a 2D CNN system, Euler method solve_cnn2d_multi- solve a 2D multi-layered CNN system, Euler method Utility routines make_cnn_params- check and update the CNN params structure make_cnn_params_multi- check and update the multi-layer CNN params structure make_cnn_time- check and update the CNN time structure step_ahead1- advance a 1D CNN state one step forward, Euler method step_ahead2- advance a 2D CNN state one step forward, Euler method unit_sat- unity saturation )

0-1整数规划有很广泛的应用背景，比如指派问题，背包问题等等，实际上TSP问题也是一个0-1问题，当然这些问题都是NP问题，对于规模较大的问题用穷举法是没有办法在可接受的时间内求得最优解的，本程序只不过是一个练习，得意之处是用递归法把所有解都排列出来。另：胡运权所著的《运筹学基础及应用（第三版）》第97页的例3，我用本程序求解得到的结果是：最优解是x*=(1,0, 0, 0, 0)，最优值是f(x*)=8，但书求得最优解是x*=(1,0, 1, 0, 0)，最优值是f(x*)=4，是不是书中写错了，请大家验证。以下是源程序，大家可以任意使用无版权问题，另外，如果大家有大规模的0-1规划的问题也希望提供给我，谢谢。变量个数至少是3个(0-1 Integer Programming is a very broad background, such as assignment, bags and so on. actually TSP is a 0-1, of course, these problems are NP, for larger problem with the law is not exhaustive approach in an acceptable time frame to obtain the optimal solution. this procedure is only one practice, farewell tour is the method used recursive all of the solutions to show. Another : Hu Yun-author of "operational research, and application base (third edition)" Article 97 of the three cases, I use this procedure to solve the result is : is the optimal solution x = (1,0, 0, 0, 0), the optimal values of f (x*) = 8, but the book is to find the optimum solution x = (1,0, 1, 0, 0), the optimal values of f (x*) = 4, the book is not a mistake, please certification. Following is the source)

用Simulink实现PAM调制，并显示星座图和眼图(PAM modulation and demodulation)

constrained evidential c means

说明：  用Jacobi和Gauss-seidel方法解方程组。并比较迭代次数，是误差小于10E-3(Use of Jacobi and Gauss-seidel method to solve equations. And compare the number of iterations is less than 10E-3 error)

伪wigner分布，强大好用。修改输入和输出就能很快地使用。(Pseudo wigner distribution, a powerful easy to use. Modify the input and output can be quickly used.)

UDF添加动量源项 fluent仿真分析虚拟螺旋桨(UDF momentum)

说明：  script code svpwm modulation

采用了一种简化的方向图模型。方位和俯仰均用高斯型。(Direction using a simplified graph model. Azimuth and elevation are used Gaussian.)