使用 selBest 选择从 DEAP 中为 eaSimple 算法选择的个体数

Question

我正在尝试运行来自 DEAP module. I would like to specify the number of individuals to be selected at each generation. However, if I specify a k parameter for the selection function 的 eaSimple 算法，但出现错误。

from deap import base, tools, creator, algorithms
import random
import numpy as np

def fitness(individual):
    x = individual[0]
    y = individual[1]
    return np.exp(-9*x*y) * np.sin(3*np.pi*x)**2 * np.sin(3*np.pi*y)**2,

toolbox = base.Toolbox()

creator.create("FitnessMax", base.Fitness, weights=(1.0,))
creator.create("Individual", list, fitness=creator.FitnessMax)
toolbox.register("individual", tools.initRepeat, creator.Individual, random.random, n=2)

toolbox.register("population", tools.initRepeat, list, toolbox.individual)

toolbox.register('evaluate', fitness)
toolbox.register('mutate', tools.mutPolynomialBounded, eta=.6, low=[0,0], up=[1,1], indpb=0.1)
toolbox.register('mate', tools.cxUniform, indpb=0.5)
toolbox.register('select', tools.selBest, k=50)

pop = toolbox.population(n=100)

pop, logbook = algorithms.eaSimple(pop, toolbox, cxpb=0.5, mutpb=0.1, ngen=100)

此示例中的最后一行引发了错误

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-2-5ee0faee2c49> in <module>
----> 1 pop, logbook = algorithms.eaSimple(pop, toolbox, cxpb=0.5, mutpb=0.1, ngen=100)

/usr/local/lib64/python3.7/site-packages/deap/algorithms.py in eaSimple(population, toolbox, cxpb, mutpb, ngen, stats, halloffame, verbose)
    163     for gen in range(1, ngen + 1):
    164         # Select the next generation individuals
--> 165         offspring = toolbox.select(population, len(population))
    166 
    167         # Vary the pool of individuals

TypeError: selBest() got multiple values for argument 'k'

请注意，将行 toolbox.register('select', tools.selBest, k=50) 替换为 toolbox.register('select', tools.selBest) 可消除错误。

此处k的默认值是多少，如何为k指定我自己的值？

Answer 1

在eaSimple算法中，实际上没有选择。从 the documentation 开始，pseudo-code 描述了幕后情况

population = select(population, len(population))

这意味着选择初始种群大小的种群子集，即整个种群进行突变、交配和评估。

然后算法做

population = offspring

所以所有由变异和交配产生的个体都取代了所有 parents。

这也（某种程度上）回答了问题的第二部分：使用 eaSimple 无法在 selBest 方法中为 k 指定自定义值。

使用 selBest 选择从 DEAP 中为 eaSimple 算法选择的个体数

Number of selected individuals for eaSimple algorithm from DEAP with selBest selection

python-3.x

deap