Proof: Suppose this algorithm works, i.e. for each position j (starting from i), the probability of any number in the range[i,n-1] to be at position j is 1/(n-i).
Let’s look at int i = random.nextInt(i, n):
(1) If j == i, nums[i] does not need to change its position, which has probability 1/(n-i).
(2) if j != i, nums[i] needs to be swapped with nums[j]. The probability of any number x in the range [i, n-1] to be at position j = nums[j] changes its position * x is at position j
= (1-1/(n-i)) * (1/n-i-1) = 1/(n-i) (选的位置是要和i交换的概率 * number x正好是选中的位置的概率)
Each number has equal probability to be at any position.
class Solution {
private int[] array;
private int[] original;
private Random rand = new Random();
private int randRange(int min, int max){
return rand.nextInt(max - min) + min;
}
private void swapAt(int i, int j){
int temp = array[i];
array[i] = array[j];
array[j] = temp;
}
public Solution(int[] nums) {
array = nums;
original = nums.clone();
}
/** Resets the array to its original configuration and return it. */
public int[] reset() {
array = original;
original = original.clone();
return original;
}
/** Returns a random shuffling of the array. */
public int[] shuffle() {
for(int i = 0;i < original.length;i++){
swapAt(i, randRange(i, original.length));
}
return array;
}
}
/**
* Your Solution object will be instantiated and called as such:
* Solution obj = new Solution(nums);
* int[] param_1 = obj.reset();
* int[] param_2 = obj.shuffle();
*/