Question

我正在尝试使用numpy对遗传算法种群进行交叉。我已使用父级1和父级2分割了人口。

population = np.random.randint(2, size=(4,8))
p1 = population[::2]
p2 = population[1::2]

但是我无法找出任何lambda或numpy命令对父母进行多点交叉。该概念是将ith row of p1并与ith row of p2 随机交换一些位。

Answer 1

我认为您想从p1和p2中随机选择，逐个单元地选择。

为了更容易理解，我将p1更改为10到15，将p2更改为20到25。在这些范围内随机生成p1和p2。

p1
Out[66]: 
array([[15, 15, 13, 14, 12, 13, 12, 12],
       [14, 11, 11, 10, 12, 12, 10, 12],
       [12, 11, 14, 15, 14, 10, 13, 10],
       [11, 12, 10, 13, 14, 13, 12, 13]])

In [67]: p2
Out[67]: 
array([[23, 25, 24, 21, 24, 20, 24, 25],
       [21, 21, 20, 20, 25, 22, 24, 22],
       [24, 22, 25, 20, 21, 22, 21, 22],
       [22, 20, 21, 22, 25, 23, 22, 21]])

In [68]: sieve=np.random.randint(2, size=(4,8))
In [69]: sieve
Out[69]: 
array([[0, 1, 0, 1, 1, 0, 1, 0],
       [1, 1, 1, 0, 0, 1, 1, 1],
       [0, 1, 1, 0, 0, 1, 1, 0],
       [0, 0, 0, 1, 1, 1, 1, 1]])
In [70]: not_sieve=sieve^1  # Complement of sieve

In [71]: pn = p1*sieve + p2*not_sieve

In [72]: pn
Out[72]:
array([[23, 15, 24, 14, 12, 20, 12, 25],
       [14, 11, 11, 20, 25, 12, 10, 12],
       [24, 11, 14, 20, 21, 10, 13, 22],
       [22, 20, 21, 13, 14, 13, 12, 13]])

当筛子为1时，青少年中的数字来自p1 筛子为0时，二十年代的数字来自p2

也许可以提高效率，但这是您期望的输出结果吗？

使用Numpy的MultiPoint交叉

1 个答案: