1.整体流程(有不懂的地方可以在评论区留言)
? 初始化种群,计算适应性,选择,交叉,变异,更新并评估,判断是否达到最大迭代次数,输出最佳目标函数
遗传算法有4个参数要提前设定,一般范围是:
群体大小:20-100
遗传算法终止进化代数:100-500
交叉概率:0.4-0.99
变异概率:0.0001-0.1
2.求9*sin(5*x)+8*cos(4*x)最优值(matlab)
clear all;
popsize=20; %群体大小
chromlength=10; %字符串长度(个体长度)
pc=0.7; %交叉概率
pm=0.005; %变异概率
pop=initpop(popsize,chromlength); %随机产生初始群体
for i=1:20 %20为迭代次数
value=calobjvalue(pop); %计算目标函数
fitvalue=calfitvalue(value); %计算群体中每个个体的适应度
newpop=selection(pop,fitvalue); %复制
newpop=crossover(pop,pc); %交叉
newpop=mutation(pop,pc); %变异
[bestindividual,bestfit]=best(pop,fitvalue);%求出群体中适应值最大的个体及其适应值
y(i)=max(bestfit);
n(i)=i;
pop3=bestindividual;
x(i)=decodechrom(pop3,1,chromlength)*10/1023;
pop=newpop;
end
fplot('9*sin(5*x)+8*cos(4*x)');
hold on;
plot(x,y,'r+');
hold off;
3.求y=sin(10*x)*x + cos(2*x)*x 在【0, 5】区间最大值(Python)
import numpy as np
DNA_SIZE = 10 # DNA length
POP_SIZE = 100 # population size
CROSS_RATE = 0.8 # mating probability (DNA crossover)
MUTATION_RATE = 0.003 # mutation probability
N_GENERATIONS = 200
X_BOUND = [0, 5] # x upper and lower bounds
def F(x):
return np.sin(10*x)*x + np.cos(2*x)*x # to find the maximum of this function
def get_fitness(pred):
return pred - np.min(pred) + 1e-3 # 防止出现非正值,因为 select 时概率非负。
def translateDNA(pop):
# np.arange(DNA_SIZE)[::-1] --> array([9, 8, 7, 6, 5, 4, 3, 2, 1, 0])
# 2 ** np.arange(DNA_SIZE)[::-1] --> array([512, 256, 128, 64, 32, 16, 8, 4, 2, 1], dtype=int32)
# pop.dot(2 ** np.arange(DNA_SIZE)[::-1]) --> 将二进制转换为十进制
# (2 ** DNA_SIZE - 1) --> 1023
# pop.dot(2 ** np.arange(DNA_SIZE)[::-1]) / (2 ** DNA_SIZE - 1) --> 计算染色体值占最大值(1023)的比例
# pop.dot(2 ** np.arange(DNA_SIZE)[::-1]) / (2 ** DNA_SIZE - 1) * X_BOUND[1] --> 将范围从[0, 1023]缩放到[0, 5]
return pop.dot(2 ** np.arange(DNA_SIZE)[::-1]) / (2 ** DNA_SIZE - 1) * X_BOUND[1]
def select(pop, fitness):
# 在 np.arange(POP_SIZE) 中依概率 p 选出 size 个染色体代号
# replace=True 表示选出的染色体可以有重复的
idx = np.random.choice(np.arange(POP_SIZE), size=POP_SIZE, replace=True, p=fitness / fitness.sum())
return pop[idx]
def crossover(parent, pop):
if np.random.rand() < CROSS_RATE:
i_ = np.random.randint(0, POP_SIZE, size=1) # 选定 pop 中需要交叉的染色体,如 i_ = 1,表示第二条染色体
# 选择需要改变的基因,如 cross_points = [False True True False False True False True True False]
# 表示第2,3,6,8,9个位置上的基因会改变
cross_points = np.random.randint(0, 2, size=DNA_SIZE).astype(np.bool)
# pop[i_, cross_points] 会得到 pop 种群中 i_ 这条染色体上的 cross_points 中 True 位置上的基因
# parent[cross_points] = pop[i_, cross_points] 将刚刚得到的基因放到 parent 这条染色体的 cross_points 中 True 的位置上
parent[cross_points] = pop[i_, cross_points]
return parent
def mutate(child):
for point in range(DNA_SIZE):
if np.random.rand() < MUTATION_RATE:
child[point] = 1 if child[point] == 0 else 0
return child
pop = np.random.randint(2, size=(POP_SIZE, DNA_SIZE)) # initialize the pop DNA
print(pop.shape)
for _ in range(N_GENERATIONS):
F_values = F(translateDNA(pop)) # compute function value by extracting DNA
# GA part (evolution)
fitness = get_fitness(F_values)
print("Most fitted DNA: ", pop[np.argmax(fitness), :], "适应度为:", round(fitness[np.argmax(fitness)], 3))
pop = select(pop, fitness)
pop_copy = pop.copy()
for parent in pop:
child = crossover(parent, pop_copy)
child = mutate(child)
parent = child # 用经过交叉变异的下一代(新的染色体)来代替上一代(pop 种群中选出的)
print('最大值为:', F(translateDNA(pop[np.argmax(fitness), :])))
|