问题
I have a 2d array, matrix of a sort (m x n). I need to generate '1' in k cells, but the probability of it should be equal for each cell.
for example, if k=3, we pick randomly where to place the 3 '1's :
[0, 0, 0, 0]
[0, 1, 1, 0]
[1, 0, 0, 0]
At first, I tackled this by generating a Random of modulu m * n (rows * columns). But, that means that we could theoretically get to the end of the matrix without generating a single '1'.
Then, I read about Yates Shuffle, but wasn't sure whether that's wise and even feasible to implement it with that.
What is an efficient way to implement this?
回答1:
This is essentially a problem of "sampling without replacement": Out of mn cells, choose k of them without replacement. There are many approaches to this problem, depending on how big mn is in relation to k. If mn is relatively small, then a Fisher–Yates shuffle will work well; make a list of cells, shuffle them, then take the first k cells to set to 1.
For more details, see my sections on sampling without replacement and shuffling. (Part of my comment moved here:) Different sampling algorithms have different tradeoffs in terms of time and space. For example, a Fisher–Yates shuffle has time and space complexity of O(mn), while a partial shuffle can have time complexity O(k).
来源：https://stackoverflow.com/questions/61065565/generating-placing-k-random-items-in-a-2d-array