// Author: Adrian Kuegel
// Date: January 16th, 2007
// Algorithm: Dynamic Programming
// Complexity: O(n^2)
#include <stdio.h>
#include <algorithm>
#include <assert.h>
#include <vector>
#include <map>
using namespace std;
#define prev asdadasd
#define INF 100000000
#define MAXN 1024
typedef vector<int> VI;
typedef pair<int,int> PII;
typedef vector<PII> VPII;
// simulates the sorting process and prints out the moves
// used specifies the numbers which keep their position
// p is the current permutation of the numbers
int simulate(char *used, VI &p) {
int n = p.size();
int cost = 0;
VPII moves;
for (int i=n-1; i>=0; i--) {
if (used[i])
continue;
// find element i
VI::iterator it = find(p.begin(),p.end(),i);
assert(it != p.end());
int from_pos = distance(p.begin(),it) + 1;
p.erase(it);
// find element i+1
it = find(p.begin(),p.end(),i+1);
// reinsert i before element i+1
p.insert(it, i);
int to_pos = distance(p.begin(),it) + 1;
cost += from_pos + to_pos;
moves.push_back(PII(from_pos,to_pos));
}
for (int i=0; i<n; i++)
assert(p[i] == i);
// print the result
printf("%d\n",(int)moves.size());
for (VPII::iterator it=moves.begin(); it!=moves.end(); ++it)
printf("%d %d\n",it->first,it->second);
return cost;
}
char used[MAXN+1];
int dp[MAXN+1][MAXN+1],prev[MAXN+1][MAXN+1];
void solve(VI &p) {
int n = p.size();
VI pos(n+1);
p.push_back(n);
for (int i=0; i<=n; i++)
pos[p[i]] = i;
for (int i=0; i<n; i++)
dp[n][i] = INF;
dp[n][n] = 0;
for (int i=n-1; i>=0; --i) {
int v = 1;
// determine the current position v, assuming all bigger elements already moved
// we can only make this assumption because we adjust the costs of not moving an
// element accordingly (see below)
for (int j=0; j<pos[i]; ++j)
if (p[j]<i)
++v;
int v2 = 1;
// move i before element i+1
// h is the old index position where it moves to
// v2 is the "real" position where it moves to
for (int h=0; h<=n; h++) {
if (p[h] < i)
++v2;
dp[i][h] = dp[i+1][h]+v+v2;
prev[i][h] = h;
}
// try to let i at its current position
// add specifies the additional costs for elements which are skipped
int add = 0;
for (int k=pos[i]+1; k<=n; k++) {
int tcost = dp[i+1][k]+add;
if (dp[i][pos[i]] > tcost) {
dp[i][pos[i]] = tcost;
prev[i][pos[i]] = k;
}
if (p[k]<i) {
// there are i - p[k] elements which will be moved before it
// since they are all bigger than p[k] they wouldn't be counted
// when determining the current position
add += (i-p[k]);
}
}
}
int mincost = INF,ind = -1;
for (int i=0; i<=n; i++)
if (dp[0][i] < mincost) {
mincost = dp[0][i];
ind = i;
}
assert(ind >= 0);
// determine which numbers kept their position
for (int i=0; i<n; ++i) {
used[i] = (ind == pos[i])?1:0;
ind = prev[i][ind];
}
assert(ind == n);
used[n] = 1;
int tcost = simulate(used,p);
assert(tcost == mincost);
}
int main() {
int n;
map<int,int> byscore;
scanf("%d",&n);
assert(2 <= n && n <= 1000); // assert valid range for n
for (int i=0; i<n; ++i) {
int val;
scanf("%d",&val);
assert(val >= 0 && val <= 1000000); // assert valid range
assert(byscore.find(val) == byscore.end()); // assert distinct scores
byscore[val] = i;
}
int pos = 0; // position counter
VI p(n); // desired final positions
for (map<int,int>::reverse_iterator it=byscore.rbegin(); it!=byscore.rend(); ++it,++pos)
p[it->second] = pos;
solve(p);
return 0;
}
Compilation message
sorting.cpp: In function 'int main()':
sorting.cpp:138:7: warning: ignoring return value of 'int scanf(const char*, ...)', declared with attribute warn_unused_result [-Wunused-result]
scanf("%d",&n);
~~~~~^~~~~~~~~
sorting.cpp:143:8: warning: ignoring return value of 'int scanf(const char*, ...)', declared with attribute warn_unused_result [-Wunused-result]
scanf("%d",&val);
~~~~~^~~~~~~~~~~
| # |
결과 |
실행 시간 |
메모리 |
Grader output |
| 1 |
Correct |
3 ms |
376 KB |
Output is correct |
| 2 |
Correct |
3 ms |
488 KB |
Output is correct |
| 3 |
Correct |
4 ms |
524 KB |
Output is correct |
| 4 |
Correct |
4 ms |
812 KB |
Output is correct |
| 5 |
Correct |
3 ms |
1368 KB |
Output is correct |
| 6 |
Correct |
4 ms |
2128 KB |
Output is correct |
| 7 |
Correct |
9 ms |
4512 KB |
Output is correct |
| 8 |
Correct |
13 ms |
6988 KB |
Output is correct |
| 9 |
Correct |
13 ms |
8652 KB |
Output is correct |
| 10 |
Correct |
14 ms |
8668 KB |
Output is correct |
