在 DEAP 中访问多个统计数据

Access multiple statistics in DEAP

the DEAP tutorial 之后,我正在尝试使用多个统计信息。

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

# Define toolbox
toolbox = base.Toolbox()

creator.create("FitnessMin", base.Fitness, weights=(-1.0,))
creator.create("Individual", list, fitness=creator.FitnessMin)
IND_SIZE=5
toolbox.register("individual", tools.initRepeat, creator.Individual, random.random, n=IND_SIZE)

toolbox.register("population", tools.initRepeat, list, toolbox.individual)
popul = toolbox.population(n=7)

def evaluate(individual):
    a = sum(individual)
    b = len(individual)
    return (a / b,)

toolbox.register('evaluate', evaluate)
toolbox.register('mutate', tools.mutGaussian, mu=0.0, sigma=0.2, indpb=0.1)
toolbox.register('mate', tools.cxUniform, indpb=0.4)
toolbox.register('select', tools.selBest)

# Define statistical tools
stats_fit = tools.Statistics(key=lambda ind: ind.fitness.values)
stats_size = tools.Statistics(key=len)
mstats = tools.MultiStatistics(fitness=stats_fit, size=stats_size)

mstats.register("avg", np.mean)
mstats.register("std", np.std)

# Run algorithm
popul, mlogbook = algorithms.eaSimple(popul, toolbox, cxpb=0.5, mutpb=0.2, ngen=10,
                                      stats=mstats, verbose=True)

这会在屏幕上打印统计数据。

文档指出

The multi-statistics object can be given to an algorithm [...] using the exact same procedure as the simple statistics.

但是当我尝试访问日志时,它 returns 是 None 的列表。

In: mlogbook.select("avg")

Out: [None, None, None, None, None, None, None, None, None, None, None]

如果我使用常规统计而不是多次统计,我不会遇到这个问题:logbook.select("avg") returns 一个浮点数列表。

如何在使用多个统计数据时访问记录的统计数据?

需要指定选择统计的章节。例如,

mlogbook.chapters["fitness"].select("avg")

returns跨代适应度平均值的期望列表。