q-learning

there is an other java project in eclipse explaine how to code a simple calculator with a motor class

滤波程序，从同事那里要的，大家一起交流学习(BADNELL & BURGESS D.A.M.T.P. CAMBRIDGE C C DIAGONALIZATION OF REAL SYMMETRIC N-BY-N MATRIX Z. C C METHOD: HOUSEHOLDER REDUCTION TO TRI-DIAGONAL FORM AND SHIFTED C QL ALGORITHM TO DETERMINE THE E-VALUES AND E-VECTORS. C C BASED ON MARTIN, REINSCH & WILKINSON: NUM. MATH. 11, 181-95 (1968). C C INPUT REQUIRED. N, IUP AND Z. ONLY LOWER TRIANGLE OF Z NEED BE SUPPLIED. C MATRIX Z OVERWRITTEN BY EIGENVECTORS OF Z. C IUP=1/-1 ASC/DESCENDING SORT, 0 NO SORT. C MXMAT, IS THE ROW DIMENSION OF Z IN THE CALLING ROUTINE. C C OUTPUT. Z AND D, WHERE Z CONSISTS OF COLUMN EIGENVECTORS C AND D CONSISTS OF CORRESPONDING EIGENVALUES. C C NOTE: E IS A WORKING ARRAY. C)

四阶龙格库塔算法，用来求解二阶偏微分方程，在转子系统的动力学方程数值求解中普遍应用(Runge Kutta algorithm)

用來計算帶通濾波器相關參數，可以看出響應圖(Band-pass filter is used to calculate the relevant parameters can be seen in Fig response)

一个利用共轭梯度法的正则化实例： 通过输入矩阵的维数，利用共轭梯度法的原理，进行正则化计算；Morozov偏差原理(A conjugate gradient method using the regularization Example: Enter the dimensions of the matrix, using the principle of conjugate gradient method for regularization calculation)

包括牛顿法，阻尼牛顿法，变尺度算法，最速下降法的优化算法程序。(Newton' s method, damped Newton method, variable metric algorithm, the steepest descent method optimization algorithm.)

说明：  用FORTRAN语言实现二维欧拉方程的求解，采用二阶MUSCL迎风格式()

MATLAB编写的非结构化网格生成程序，不算很长(MATLAB program written in unstructured mesh generation, not very long)

对数据进得自回归分析预测。并生成下一周期的预测值(Of data into the prediction derived from the regression analysis. And generate the next cycle of prediction value)

介绍线性算子的基本概念的文章,这是书里面的一个章节(Foundation of linear operator theory)