从 matlab 转换为 python/numpy/ 的一般问题

Question

我正在尝试 'translate' 使用 numpy python 的工作 matlab 脚本。

在matlab代码中生成了如下的某种变量：

GA.Ng=2;       % number of genes
GA.Np=Np;      % size of population
GA.NG=NG;      % number of generations
GA.pc=0.5;     % probability of crossover
GA.alpha=0.5;  % blend ratio for crossover
GA.pm=0.1;     % probability of a gene being mutated
GA.xmn=[0 0];  % vector of minimum values for unnormalized genes
GA.xmx=[5 5];  % vector of maximum values for unnormalized genes

如何在 python 中实现此目的？我试过了，但没有成功：

def example1p6A(NG, Np, rf, pf):
    GA = np.zeros(1, dtype = [('Ng', int),
                              ('Np', int),
                              ('NG', int),
                              ('pc', int),                              
                              ('alpha', float),
                              ('pm', int),
                              ('xmin', float),
                              ('xmax', float)])

    GA['Ng'] = 2                    # Number of genes
    GA['Np'] = Np                   # size of population
    GA['NG'] = NG                   # number of generations
    GA['pc'] = 0.5                  # probability of crossover
    GA['alpha'] = 0.5               # blend ratio for crossover
    GA['pm'] = 0.1                  # probability of a gene being mutated
    GA['xmin'] = np.array([0, 0])   # vector of minimum values for unnormalised genes
    GA['xmax'] = np.array([5, 5])   # vector of maximum values for unnormalised genes

    # Init population:
    P = np.random.rand(5,5)

    #return (GA['Ng'][0], Np, rf, pf)
    return P

我收到错误消息

ValueError: could not broadcast input array from shape (2) into shape (1)

Answer 1

在 Python 中，您可以使用 dictionary:

def example1p6A(NG, Np, rf, pf):
    GA = dict(Ng=2,
              Np=Np,
              NG=NG,
              pc=0.5,
              alpha=0.5,
              pm=0.1,
              xmn=[0, 0],
              xmx=[5, 5])

    P = np.random.rand(5,5)

    return (GA['Ng'][0], Np, rf, pf)

Answer 2

问题是您将 xmin 和 xmax 定义为 float，但您试图将它们分配为数组。所以这就是你收到错误的原因。您正在尝试将“来自形状 (2) 的输入数组”分配给具有“形状 (1)”的对象。所以，解决方案是将xmin和xmax定义为float的数组。下面是一个可以使其正常工作的示例。

def example1p6A(NG, Np, rf, pf):
    GA = np.zeros(1, dtype = [('Ng', int),
                              ('Np', int),
                              ('NG', int),
                              ('pc', int),                              
                              ('alpha', float),
                              ('pm', int),
                              ('xmin', (float, (2,))),
                              ('xmax', (float, (2,)))])

    GA['Ng'] = 2                    # Number of genes
    GA['Np'] = Np                   # size of population
    GA['NG'] = NG                   # number of generations
    GA['pc'] = 0.5                  # probability of crossover
    GA['alpha'] = 0.5               # blend ratio for crossover
    GA['pm'] = 0.1                  # probability of a gene being mutated
    GA['xmin'] = np.array([0, 0])   # vector of minimum values for unnormalised genes
    GA['xmax'] = np.array([5, 5])   # vector of maximum values for unnormalised genes

    # Init population:
    P = np.random.rand(5,5)

    #return (GA['Ng'][0], Np, rf, pf)
    return P

有关这方面的更多信息，请查看此 link:

https://numpy.org/doc/stable/reference/arrays.dtypes.html

从 matlab 转换为 python/numpy/ 的一般问题

General issues translating from matlab to python/numpy/

python

matlab

numpy