一个简单的遗传算法

import numpy as np

size=10
UPPER_BOUND=4
LOWER_BOUND=0
deviation=0.25
num_offspring=3
num_iter=30

def check_bound(x):
    if x<LOWER_BOUND:
        return LOWER_BOUND
    if x>UPPER_BOUND:
        return UPPER_BOUND
    return x

def fitness(x):
    return -x*(x-1)*(x-2)*(x-3)*(x-4)

def get_birth(x,y):
    child = (x+y)*0.5
    child = child + np.random.normal(0,deviation)
    return check_bound(child)

#init
population = np.random.uniform(LOWER_BOUND,UPPER_BOUND,size)

for iter in range(num_iter):
  population.sort(key=lambda x:fitness(x))

  #offspring
  offspring=[]
  mother=population[-num_offspring*2::2]
  father=population[-num_offspring*2+1::2]
  offspring=[]
  for x,y in zip(mother,father)
    offspring.append(get_birth(x,y))
 
  population=population.extend(offspring)
  #die
  population.sort(key=lambda x:fitness(x))
  population=population[-size:]
  print(population)