python - 如何使用 Pyevolve 在实值染色体内设置不同的参数范围？

Question

我正在尝试使用 pyevolve 来实现实值遗传算法。 (此处给出示例文档:http://pyevolve.sourceforge.net/examples.html#example-2-real-numbers-gaussian-mutator)

可以使用setParams设置参数的范围(本例中为20)，如下所示:

# 基因组实例
基因组 = G1DList.G1DList(20)
基因组.setParams(rangemin=-6.0, rangemax=6.0)

但是，相同的范围适用于所有 20 个。我希望参数有不同的范围。我尝试的方法是更改​​初始化器文件。

文件中原始的相关部分是:

    def G1DListInitializatorReal(genome, **args):
       """ Real initialization function of G1DList

        This initializator accepts the *rangemin* and *rangemax* genome parameters.

       """
       genome.clearList()

       for i in xrange(genome.listSize):
          randomReal = rand_uniform(genome.getParam("rangein", 0),
                            genome.getParam("rangemax", 100))
          genome.append(randomReal)

我的修改(假设前 15 个有一个范围，后 5 个有另一个范围)是这样的:

     def G1DListInitializatorReal(genome, **args):

       genome.clearList()

       for i in xrange(0,15):
          print i
          randomReal = rand_uniform(genome.getParam("rangein_1", 0),
                            genome.getParam("rangemax_1", 100))
          genome.append(randomReal)
       for j in xrange(15,20):
          print j
          randomReal2 = rand_uniform(genome.getParam("rangein_2", 0),
                            genome.getParam("rangemax_2", 100))
          genome.append(randomReal2)

我添加了索引 i 和 j 的打印，以确保我知道这个索引正在被调用。我已将修改后的 Initializators 文件放在与我的代码相同的文件夹中，但当我运行它时，它会从其他地方调用原始文件。我觉得我错过了需要在 pyevolve 中进行的更多更改，或者我没有正确调用 Initializators，或者......我不知道。

如何成功更改 pyevolve 中染色体参数的范围？

提前致谢。

Answer 1

您可以使用等位基因，请参阅此 example .

创建等位基因:

alleles = GAllele.GAlleles()

为每个参数创建范围:

alleles.add(GAllele.GAlleleRange(range_min, range_max))

设置参数:

genome.setParams(alleles)

以下代码使用名为“data”的列表来定义范围。

from pyevolve import GAllele
from pyevolve import G1DList
from pyevolve import GSimpleGA
from pyevolve import Crossovers
from pyevolve import Initializators
from pyevolve import Mutators
from pyevolve import Scaling

# data values are the [min, max] for each parameter
data=[ [0,99], [1,10], [1,10], [90,110], [1,10],
[1,10], [1,10], [1,10], [1,10], [1,10],
[1,10], [5,10], [1,10], [1,10], [50,100],
[50,100], [50,100], [50,100], [50,100], [7,77] ]

def Grid_Constructor(a=data):
    alleles = GAllele.GAlleles()
    for i in range(0, 20):
        alleles.add(GAllele.GAlleleRange(a[i][0], a[i][1], real=True))
    return alleles

# set the params, initializator and mutator
genome = G1DList.G1DList(20)
genome.setParams(allele=Grid_Constructor())
genome.initializator.set(Initializators.G1DListInitializatorAllele)
genome.mutator.set(Mutators.G1DListMutatorAllele)
ga = GSimpleGA.GSimpleGA(genome)
print(Grid_Constructor())
#ga.evolve(freq_stats=10)

此代码的输出是:

- GAlleles
    Homogeneous:     False
    List size:   20
    Alleles:

Allele for 0 position:
- GAlleleRange
    Real:        True
    Ranges Count:    1
    Range List:
             Range from [0] to [99]

Allele for 1 position:
- GAlleleRange
    Real:        True
    Ranges Count:    1
    Range List:
             Range from [1] to [10]

Allele for 2 position:
- GAlleleRange
    Real:        True
    Ranges Count:    1
    Range List:
             Range from [1] to [10]

.
.
.

Allele for 19 position:
- GAlleleRange
    Real:        True
    Ranges Count:    1
    Range List:
             Range from [7] to [77]




------------------
(program exited with code: 0)
Press return to continue