循环排序是一种就地,不稳定的排序算法,这是一种比较排序,在理论上,对原始数组的总写入数是最佳的,这不同于其他任何就地排序算法。基于这样的思想,可以将要排序的排列分解为多个周期,可以将其单独旋转以给出排序的结果。
与几乎所有其他类型不同,项目永远不会被写入数组中的其他地方,仅仅是为了将它们推到行动之外。如果每个值已经在正确的位置,则将其写入零次,或者将一次写入其正确的位置。这与完成就地排序所需的最少重写次数相匹配。
当对一些庞大的数据集进行写操作非常昂贵时,例如在诸如EEPROM之类的EEPROM中,每次写操作都会减少存储器的使用寿命,因此将写操作的次数减到最少很有用。
Input: a[]={7,4,3,5,2,1,6}
Output: 1 2 3 4 5 6 7arr[] = {10, 5, 2, 3}
index = 0 1 2 3
cycle_start = 0
item = 10 = arr[0]
Find position where we put the item,
pos = cycle_start
while (arr[i] < item)
pos++;
We put 10 at arr[3] and change item to
old value of arr[3].
arr[] = {10, 5, 2, 10}
item = 3
Again rotate rest cycle that start with index '0'
Find position where we put the item = 3
we swap item with element at arr[1] now
arr[] = {10, 3, 2, 10}
item = 5
Again rotate rest cycle that start with index '0' and item = 5
we swap item with element at arr[2].
arr[] = {10, 3, 5, 10 }
item = 2
Again rotate rest cycle that start with index '0' and item = 2
arr[] = {2, 3, 5, 10}
Above is one iteration for cycle_stat = 0.
Repeat above steps for cycle_start = 1, 2, ..n-2#include<iostream>
using namespace std;
void cycleSort(int a[], int n) {
int writes = 0;
for (int c_start = 0; c_start <= n - 2; c_start++) {
int item = a[c_start];
int pos = c_start;
for (int i = c_start + 1; i < n; i++)
if (a[i] < item)
pos++;
if (pos == c_start)
continue;
while (item == a[pos])
pos += 1;
if (pos != c_start) {
swap(item, a[pos]);
writes++;
}
while (pos != c_start) {
pos = c_start;
for (int i = c_start + 1; i < n; i++)
if (a[i] < item)
pos += 1;
while (item == a[pos])
pos += 1;
if (item != a[pos]) {
swap(item, a[pos]);
writes++;
}
}
}
}
int main() {
int a[] ={7,4,3,5,2,1,6};
int n = 7;
cycleSort(a, n);
for (int i = 0; i < n; i++)
cout << a[i] << " ";
return 0;
}