用Java实现的多层BP神经网络，用于颜色和字符识别

改进的BP算法程序,希望对大家有所帮助.如果需要，请-The BP algorithm to improve the process, everyone would like to help. If necessary, please take a look at.

遗传算法的源程序，对于初学者有很大帮助，建议多-Genetic algorithm source code, are very helpful for beginners, the proposed multi-

各种数值算法，包括：二叉树、贪心算法、搜索算法、遗传算法、模拟退火算法、分治策略等等。-A variety of numerical algorithms, including: binary tree, greedy algorithm, search algorithm, genetic algorithm, simulated annealing algorithm, sub-rule tactics and so on.

模拟退火算法 模拟退火算法(Simulated Annealing,简称SA算法)是模拟加热熔化的金属的退火过程，来寻找全局最优解的有效方法之一。 模拟退火的基本思想和步骤如下： 设S＝{s1,s2,…,sn}为所有可能的状态所构成的集合， f：S―R为非负代价函数，即优化问题抽象如下： 寻找s*∈S，使得f(s*)＝min f(si) 任意si∈S （1）给定一较高初始温度T，随机产生初始状态S （2）按一定方式，对当前状态作随机扰动，产生一个新的状态S’ S’＝S＋sign(η).δ 其中δ为给定的步长， η为[－1,1]的随机数-simulated annealing algorithm (Simulated Annealing, or SA algorithm) is a simulation of heating molten metal in the annealing process, to find the global optimum one of the effective ways. Simulated Annealing basic ideas and the steps are as follows : S = (s1, s2, ..., sn) for all possible state posed by the pool, f : S-R non-negative cost function, that is abstract optimization problems are as follows : Find S* s, making f (s*) = min f (si) arbitrary si S (1) to set a higher initial temperature T, randomly generated initial state S (2) of a certain form, the current state of random disturbance, have a new state S "S" = S+ sign (). delta where given for the step, [-1,1] Random Number

本例是本人写的遗传算法的最简单的一个实现，并且带有一个说明文档。用C语言编写，在TurborC中调试，由于当时在随机数产生上没有种种子，所以极易收敛，有兴趣可以种上种子试一下。-the cases I wrote is the most genetic algorithm to achieve a simple, and with a documented. C language, in which TurborC debugging, because the random number generator not all sons, so vulnerable convergence are interested can plant seeds try.

加了动量项和域值，比较好用的bp网络算法-added value items and Victoria, better quality bp Network

人工智能领域全新的研究结果-new areas of artificial intelligence research

神经网络 MCP运算法-neural network algorithm

用BP神经网络逼近函数Z=SIN(X*Y),采用C++编程.-Using BP neural network to approach the function Z = SIN (X* Y), using C++ programming.

神经网络与遗传算法结合，需要的朋友就不用多说了，仅供学习和参考。-Neural network and genetic algorithm combined with the need for friends goes without saying, for learning and reference.