遗传算法(C#)

遗传算法

示例为计算函数为y = -x*x+ 5的最大值，-32<=x<=31.

参考: http://blog.csdn.net/ljp1919/article/details/42425281

采用轮盘赌算法作为选择算子的算法

核心算法

轮盘赌

        /// <summary>        /// 更新下一代;        /// 基于轮盘赌选择方法，进行基因型的选择;        /// </summary>        static void UpdateNext()        {            // 获取总的fit;            double totalFitValue = 0;            for (int i = 0; i < chromosomes.Count; i++)            {                //适应度为负数的取0;                if (chromosomes[i].fitValue <= 0)                {                    totalFitValue += 0;                }                else                {                    totalFitValue += chromosomes[i].fitValue;                }            }            Console.WriteLine("totalFitValue " + totalFitValue);            //算出每个的fit percent;            for (int i = 0; i < chromosomes.Count; i++)            {                if (chromosomes[i].fitValue <= 0)                {                    chromosomes[i].fitValuePercent = 0;                }                else                {                    chromosomes[i].fitValuePercent = chromosomes[i].fitValue / totalFitValue;                }                Console.WriteLine("fitValuePercent " + i + " " + chromosomes[i].fitValuePercent);            }            //计算累积概率;            // 第一个的累计概率就是自己的概率;            chromosomes[0].probability = chromosomes[0].fitValuePercent;            for (int i = 1; i < chromosomes.Count; i++)            {                // 上一个的累计概率加上自己的概率,得到自己的累计概率;                chromosomes[i].probability = chromosomes[i - 1].probability + chromosomes[i].fitValuePercent;            }            //轮盘赌选择方法,用于选出前两个;            for (int i = 0; i < chromosomes.Count; i++)            {                //产生0-1之前的随机数;                int seed = i * 100 / 3;                double rand = new Random(seed).NextDouble();//0.0-1.0                Console.WriteLine("挑选的rand " + rand);                if (rand < chromosomes[0].probability)                {                    chromosomes[i] = chromosomes[0];                }                else                {                    for (int j = 0; j < chromosomes.Count - 1; j++)                    {                        if (chromosomes[j].probability <= rand && rand <= chromosomes[j + 1].probability)                        {                            chromosomes[i] = chromosomes[j + 1];                        }                    }                }            }        }

交叉操作

        /// <summary>        /// 交叉操作;        /// </summary>        static void CrossOperate()        {            /**         bit[5]~bit[0]   fit             * 4        000 110         12             * 3        001 010         9             * child1   000 010         14             * child2   001 110         5             */            int rand = new Random().Next(0, 6);//0-5;            Console.WriteLine("交叉的rand " + rand);            for (int i = 0; i < rand; i++)            {                //将第0个给第2个;                chromosomes[2].bits[i] = chromosomes[0].bits[i];//第一条和第三条交叉;                chromosomes[3].bits[i] = chromosomes[1].bits[i];//第二条和第四条交叉;            }            for (int i = rand; i < 6; i++)            {                chromosomes[2].bits[i] = chromosomes[1].bits[i];//第一条和第三条交叉;                chromosomes[3].bits[i] = chromosomes[0].bits[i];//第二条和第四条交叉;            }        }

变异操作

        /// <summary>        /// 变异操作;        /// </summary>        static void VariationOperate()        {            int rand = new Random(DateTime.Now.Millisecond).Next(0, 50);            Console.WriteLine("变异的rand " + rand);            if (rand == 25)//1/50 = 0.02的概率进行变异;rand==25;            {                Console.WriteLine("开始变异");                int col = new Random(DateTime.Now.Millisecond * 10).Next(0, 6);                int row = new Random(DateTime.Now.Millisecond * 10).Next(0, 4);                // 0变为1,1变为0;                if (chromosomes[row].bits[col] == 0)                {                    chromosomes[row].bits[col] = 1;                }                else                {                    chromosomes[row].bits[col] = 0;                }            }        }

源码: https://github.com/onelei/Algorithm/tree/master/GeneticAlgorithm