使序列有序的最少交换次数(minimum swaps)
交换相邻两数
如果只是交换相邻两数,那么最少交换次数为该序列的逆序数。
交换任意两数
数字的总个数减去循环节的个数??
A cycle is a set of elements, each of which is in the place of another. So in example sequences { 2, 1, 4, 3}, there are two cycles: {1, 2} and {3, 4}. 1 is in the place where 2 needs to Go, and 2 is in the place where 1 needs to go 1, so they are a cycle; likewise with 3 and 4.
The sequences {3, 2, 1, 5, 6, 8, 4, 7 }also has two cycles: 3 is in the place of 1 which is in the place of 3, so {1, 3} is a cycle; 2 is in its proper place; and 4 is in the place of 7 which is in the place of 8 in place of 6 in place of 5 in place of 4, so {4, 7, 8, 6, 5} is a cycle. There are seven elements out of place in two cycles, so five swaps are needed.
实现:
e.g. { 2, 3, 1, 5, 6, 4}
231564 -> 6 mismatch
two cycles -> 123 and 456
swap 1,2 then 2,3, then 4,5 then 5,6 -> 4 swaps to sort
Probably the easiest algorithm would be to use a bitarray. Initialize it to 0, then start at the first 0.
Swap the number there to the right place and put a 1 there.
Continue until the current place holds the right number.
Then move on to the next 0
有序列5,4,3,2,1。共5个数。
nums [0] [1] [2] [3] [4]
5 4 3 2 1
按升序排列之后为
nums1 [0] [1] [2] [3] [4]
1 2 3 4 5
5应该到1处,1应该到5处,形成了一个循环,所以可以将它们抽象成一个环,环内换序就可以了。(这种环称为循环节)
如果把它们两个看成整体,对于整个序列来说它们占据了排好序后5,1应该在的位置,所以对于整个序列来说是有序的,
它们只是自身内部无序而已。
上例中3在原本就在的位置,可以看成一个元素的循环节。
我们可以推断出有一个循环节,就可以少交换一次,因为n个元素的循环节,只需交换n-1次即可有序。
那么对于整个序列来说,最少交换次数为 元素总数-循环节个数。
Example:
231564
000000
-> swap 2,3
321564
010000
-> swap 3,1
123564
111000
-> continue at next 0; swap 5,6
123654
111010
-> swap 6,4
123456
111111
-> bitarray is all 1's, so we're done.
#include "bits/stdc++.h" using namespace std; /* * 交换任意两数的本质是改变了元素位置, * 故建立元素与其目标状态应放置位置的映射关系 */ int getMinSwaps(vector<int> &A) { // 排序 vector<int> B(A); sort(B.begin(), B.end()); map<int, int> m; int len = (int)A.size(); for (int i = 0; i < len; i++) { m[B[i]] = i; // 建立每个元素与其应放位置的映射关系 } int loops = 0; // 循环节个数 vector<bool> flag(len, false); // 找出循环节的个数 for (int i = 0; i < len; i++) { if (!flag[i]) { int j = i; while (!flag[j]) { flag[j] = true; j = m[A[j]]; // 原序列中j位置的元素在有序序列中的位置 } loops++; } } return len - loops; } vector<int> nums; int main() { nums.push_back(1); nums.push_back(2); nums.push_back(4); nums.push_back(3); nums.push_back(5); int res = getMinSwaps(nums); cout << res << '\n'; return 0; }
交换任意区间
??
#include "bits/stdc++.h" using namespace std; /* * 默认目标映射关系是 key 1 => val 1 …… key n => val n * 如果序列不是 1~n 可以通过 map 建立新的目标映射关系 * 交换任意区间的本质是改变了元素的后继,故建立元素与其初始状态后继的映射关系 */ const int MAXN = 30; int n; int vis[MAXN]; int A[MAXN], B[MAXN]; int getMinSwaps() { memset(vis, 0, sizeof(vis)); for (int i = 1; i <= n; i++) { B[A[i]] = A[i % n + 1]; } for (int i = 1; i <= n; i++) { B[i] = (B[i] - 2 + n) % n + 1; } int cnt = n; for (int i = 1; i <= n; i++) { if (vis[i]) { continue; } vis[i] = 1; cnt--; for (int j = B[i]; j != i; j = B[j]) { vis[j] = 1; } } return cnt; } int main() { cin >> n; for (int i = 1; i <= n; i++) { cin >> A[i]; } int res = getMinSwaps(); cout << res << '\n'; return 0; }