s_p52

严格按照BP网络计算公式来设计的一个matlab程序,对BP网络进行了优化设计 优化1：设计了yyy，即在o(k)计算公式时，当网络进入平坦区时(<0.0001)学习率加大，出来后学习率又还原 优化2：v(i,j)=v(i,j)+deltv(i,j)+a*dv(i,j) (BP network in strict accordance with the formula to design a matlab program, the optimized BP network design optimization 1: Design of the yyy, that is, o (k) formula, when the network into the flat areas (< 0.0001) increase learning rate large, after the optimization of learning rate reduction 2: v (i, j) = v (i, j)+ deltv (i, j)+ a* dv (i, j) )

说明：  C++源代码，递归实现。原问题描述：鼓上蚤时迁到一个山洞里去盗宝，里面有n件宝贝，价值分别是a1,a2......an，并且满足ai大于a1+a2+....+a(i-1)的和。他身上有一个背包，能装的东西最大价值是K，他只能进洞一次，请问他最多能带出多少？ 要求：输入K，n，a1，a2......an 输出：他能最多带出的宝物价值 (C++ Source code Recursive achieve. The original problem description: drum when flea盗宝going to move in a cave, there are n pieces of treasure, valued are a1, a2 ...... an, and to meet ai than a1+ A22B !....+ a (i-1) and. Him to have a backpack, things can hold the greatest value is K, he could only hole once, to ask him to bring out the maximum number of? Requirements: Enter K, n, a1, a2 ...... an output: he can bring out the maximum value of the treasures)

关于PID的比例因子调节的文章，非常不错，对各种比例关系进行对比说明。(PID adjustment on the proportion factor of articles, very good, compared to the ratio between the various instructions.)

分别使用传统卡尔曼滤波和改进的自适应卡尔曼滤波对运动物体进行跟踪，比较并分析结果，验证自适应卡尔玛滤波的优点。(Respectively using conventional Kalman filter and adaptive Kalman filter for improved tracking of moving objects, compare and analyze the results to verify the advantages of adaptive filtering Kalmar.)

相场模拟晶枝生长，采用有限差分法计算，可更改参数获得不同结果(Phase field simulated dendrite growth)

含有噪声的多维非线性系统辨识！！！ (Noisy multi-dimensional non-linear system identification! ! !)

V-BLAST MIMO STRUCTURE

介绍怎样制作机器人，包括硬件制作 系统建模　控制算法()

ML-KNN实现过程，包括训练和测试两个过程(the process of ML-KNN, including training and testing )

说明：  经典Runge-Kutta法，计算积分的源代码，用matlab实现(Classical Runge-Kutta method to calculate the source code of points, using matlab implementation)