如何创建一个numpy矩阵，其值满足几个条件？

Question

I want to create a matrix, that meets this condition at first:

1. eg if a value m[2][1] == 0 <-> m[1][2] in {0,1}

For this one I use an upper triangular matrix like:

m = np.ones((6, 6), int)
m = np.triu(m, 1)

And I change some random values of the remaining to zero, until a certain percentage is achieved:

MaxPer = 0.75   
i, e = np.count_nonzero(m), np.count_nonzero(m)
MAX = np.round(MaxPre*e)
while i > MAX:
   m[np.random.randint(len(m[0][:]))][np.random.randint(len(m[0][:]))] = 0
   i = np.count_nonzero(m)

Now the second criteria is the one I am struggling with:

2. if m[1][2] == 1 & m[1][3] == 0 <-> m[2][3] == 0

I want to check, if all values in m satisfy this condition (indices 1,2,3 are examples). If not, I want to change the value of m[2][3] to zero. I tried the following code with for-loops, but it overwrites all the values to zero.

for k in range(len(m[0][:])):
            for l in range(k, len(m[0][:])):
                for j in range(l,len(m[0][:])):
                    if matrix[k][l] == 1 & matrix[k][j] == 0:
                        matrix[l][j] = 0

Is there a simple way to this without the loops?

Answer 1

I think imposing your second condition indeed causes zeroing of the matrix. Are you sure you want

m[i][j] == 1 & m[i][k] == 0 <-> m[j][k] == 0

without any dependencies between i, j, and k?

Maybe you want something like this (which would require a few changes in the fors)?

m[i][i+1] == 1 & m[i][i+2] == 0 <-> m[i+1][i+2] == 0