SVM

考虑单独变量之间的相关性，将其变成扩展变量，再用留一法进行数据处理，最后用SVM进行回归(Consider the correlation between individual variables, to expand into the variable, and then leave one for data processing, and finally with SVM regression)

The function takes at least three input arguments (function, its derivative, and an initial guess). After iterations, it gives the root of the function. The fourth input argument is to change the stopping criteria.

Solve differential equations, higher order method.DFSOLVE integrates a single ordinary differential equation using 4th and 5th order Runge-Kutta formulas. This routine is designed to be used only with DFIELD.

Implemenation of lms algorithm

说明：  压缩包内自带飞行数据，直接可用，附带结果说明word(The flight data is provided in the compressed package, which is directly available, and the result description word is attached)

模糊pid控制，并已经通过调试，程序可行。非常有用(Fuzzy pid control, and has been through the debugging, the program is feasible. Very useful)

采用Hamming窗实现的低通滤波器，可设置通带频率、截止频率以及带宽等滤波器特性参数。(The low-pass filter is realized by using Hamming window, can set the passband filter characteristic parameters such as frequency, the cut-off frequency and bandwidth.)

查询公交，查询公交，查询公交，查询公交，查询公交，查询公交(Search bus)

(1)针对在线计算量大这一缺陷，将预测控制中的柔化输出信号的思想推广到柔化输入信号，使得约束条件被简化为仅对当前控制量的约束，可以直接计算得出；同时该方法避免了求逆矩阵，大大减小了计算量，并能够保证控制算法的可行性和良好的控制性能。 (2)针对传统算法中设计参数整定困难这一缺点，应用基于BP神经网络变参数设计的广义预测控制算法，实现了对控制量柔化参数的在线调整。 (3)利用带有遗忘因子的最小二乘法对系统辨识。本文通过仿真发现该方法对于Hénon混沌系统并不完全适用，可考虑利用其他优化系统辨识的方法对本方法进行改进，以期达到更好的辨识效果。 (4)针对系统稳定性分析复杂，本文在控制增量前加入前馈因子，保证所选的Lyapunov函数使闭环系统满足Lyapunov稳定判据，由此证明闭环系统稳定。 (1. To solve the problem of GPC huge computation, algorithm with input increment constraints is presented in which the concept of output softness was used to soften the input increments．As a result, the constraints are simplified to be the only one constraint on the current control increment which can be computed directly. At the same time, it needn’t computing the inverse matrix and thus reduces large computation. Moreover, it guarantees the feasibility of the algorithm and has good control performance． 2. To overcome the difficulty in the choice of tuning parameters in traditional GPC, a GPC algorithm with variable parameter design based on BP neural network. is presented，in which the input softness parameters are tuned on line. 3. In this paper, we Identify system by using the least square method with forgetting factor. However, after system simulation, we realize that this method doesn’t fit the Hénon chaotic system perfectly. So we recommend modify this method by other Optimizati)

蒙特卡洛模拟，期权定价分析，收益模型仿真。Monte Carlo simulation for returns(Monte Carlo simulation for returns)