Q-learning

宽带信号 DOA估计 TCT算法 此算法可以较好的完成宽带信号的波达角估计 可以解相干信号 属于子空间算法(Wideband signal DOA estimates TCT algorithm this algorithm can better complete the broadband signal can be solution-of-arrival estimation of coherent signals are sub-space algorithm)

瑞利信道仿真，用于ofdm，mimo系统中，望大家下载(Rayleigh channel simulation for ofdm, mimo system, hope you download)

this is a 2d fdtd in matlab

用matlab实现蚁群算法，求解TSP以及两地最优问题，在原始的基础上改进，效果良好。(use Matlab to realize ant colony algorithm to solve TSP and optimal problem, on the basis of the original improvement, the result is good. )

检测矩形 识别圆，矩形，三角形形状 直接在图像上显示出结果(Rectangle detection Detect the rectangle recognition circle, rectangle and triangle shape, and display the results directly on the image.)

The toolbox solves a variety of approximate modeling problems for linear static models. The model can be parameterized in kernel, image, or input/output form and the approximation criterion, called misfit, is a weighted norm between the given data and data that is consistent with the model. There are three main classes of functions in the toolbox: transformation functions, misfit computation functions, and approximation functions. The approximation functions derive an approximate model from data, the misfit computation functions are used for validation and comparison of models, and the transformation functions are used for deriving one model representation from another. KEYWORDS: Total least squares, generalized total least squares, software implementation.

33节点的配电网重构程序，基于粒子群算法。相当快速。但是需要再修改源码才可以重新生成数据。(33 node distribution network reconfiguration procedures, based on particle swarm optimization. Pretty fast. But the need to modify the source code before they can re-generate the data.)

所谓阶梯式控制，就是对未来控制量的变化加以约束，强迫未来控制量以一阶指数响应曲线的形式变化。在这种控制策略下，可以避免求解控制律时的矩阵求逆，且在处理约束限制时，可方便将 和 的约束转化为对 的约束，减少了算法计算量。(The so-called ladder-type control, is the change in the future to control the amount of bound, forced to the future control the amount of the first order exponential response curve in the form of change. In this control strategy, the control law can be avoided when solving the matrix inversion, and the limit in dealing with constraints, and constraints can easily be translated into constraints, reducing the computational algorithm.)

file program berkeleye university

方向键←→控制方块的左右移动，方向键↑变形，方向键↓加速方块落下的速度(The direction key vs. control box and move around, the direction key up key down the direction of deformation, accelerate the speed of falling blocks)