def main():
average_incomes = []
populations = []
# 用于判断全局收敛的方差最小值,小于该值时认为种群达到全局收敛
std_limition = 0.1
currentAdjustGeneration = 41
Standard.init_data()
population = Standard.initialPopulation
generation = 0
time3 = datetime.now()
# 十万代以内，不停迭代，直到某一代出现完全收敛
while generation < Standard.maxgeneration:
populations.append(population)
print("第")
print(generation)
print("代")
if (generation == 40):
Standard.retainEliteNumbers = 1
print("本代精英个数:", Standard.retainEliteNumbers)
print("本代变异率:", Standard.mutationprob)
# 每相隔30代 加一次精英
if (generation - currentAdjustGeneration > 30
and Standard.retainEliteNumbers < 5):
print("下面加精英")
Standard.retainEliteNumbers += 1
currentAdjustGeneration = generation
# 对种群解码得到表现形
decode = Generation.getDecode(population, Standard)
# 得到适应度值和累计概率值
fitnessValue, cum_proba, incomes, retain_population, elite_population, elite_incomes, generation_best_chrome = Generation.getFitnessValue(
population, decode, Standard)
# 选择新的种群,并计算平均收益
newpopulations, average_income = Generation.selectNewPopulation(
retain_population, population, cum_proba, fitnessValue,
incomes, elite_population, elite_incomes, Standard)
print("本代平均收益:", average_income)
average_incomes.append(average_income)
# 如果本代方差小于一定值，则认为达到了全局收敛，迭代终止
if (Standard.stdarr[generation] < std_limition):
break
# 新种群交叉
crossPopulations = Generation.crossNewPopulation(
newpopulations, Standard.prob)
# 变异操作
mutationpopulation = Generation.mutation(crossPopulations,
Standard.mutationprob)
# 将父母和子女合并为新的种群
population = np.vstack((newpopulations, mutationpopulation))
generation += 1
print("本代演化结束,当前平均收益数组:", average_incomes)
time4 = datetime.now()
print("本代耗时:", time4 - time3)
time3 = time4
# 获取最后要输出的平均收益和各项参数
final_average_income = average_income
best_chrome = generation_best_chrome
"""
print(average_incomes)
x = [i for i in range(generation)]
y = [average_incomes[i] for i in range(generation)]
plt.plot(x, y)
plt.show()
"""
# 返回最终平均收益,以及最后一代的最优染色体
best_chrome_dealed = best_chrome[0]
for index, item in enumerate(best_chrome_dealed):
if (index in Standard.intgerIndex):
best_chrome_dealed[index] = np.ceil(item)
return final_average_income, best_chrome_dealed, generation
