今天读到一篇文章:
http://www.topcoder.com/tc?module=Static&d1=tutorials&d2=importance_of_algorithms
其中提到用随机算法找数组中第k大数字,于是试着写了一下:
#include
<
vector
>
#include < set >
using namespace std;
#include < cstdlib >
#include < cassert >
size_t rand(size_t maxn)
{
size_t d = RAND_MAX / maxn;
size_t nd = d * maxn;
size_t r;
do r = rand(); while(r >= nd);
return r % maxn;
}
int rand( int a, int b)
{
assert(a < b);
return (int)rand(b - a) + a;
}
template < typename T >
size_t kthElement( const vector < T > & v, size_t k)
{
vector<size_t> s(v.size());
for (size_t i = 0; i < v.size(); i++)
s[i] = i;
size_t curK = k;
size_t t;
for(;;)
{
size_t c = 0;
t = rand(s.size());
const T &value = v[s[t]];
for (vector<size_t>::const_iterator it = s.begin(); it != s.end(); ++it)
{
if (v[*it] < value)
c++;
}
if (c == curK)
break;
swap(s[t], s.back());
s.pop_back();
// now, the v[s[t]] is c-th element of the rest v
if (c > curK)
{
// target is smaller than v[s[t]]
// remove greater values from s
// s.erase(remove_if(s.begin(), s.end(), bind2st(greater<T>(), value)), s.end());
for (vector<size_t>::iterator it = s.begin(); it != s.end();)
{
if (v[*it] > value)
swap(*it, s.back()), s.pop_back();
else
++it;
}
}
else if (c < curK)
{
// target is greater than v[s[t]]
// remove smaller values from s
for (vector<size_t>::iterator it = s.begin(); it != s.end();)
{
if (v[*it] < value)
swap(*it, s.back()), s.pop_back();
else
++it;
}
curK -= c + 1;
}
else
break;
}
return s[t];
}
template < typename T >
T median( const vector < T > & v)
{
size_t s = v.size();
if (s & 1)
{
return v[kthElement(v, s / 2)];
}
T a = v[kthElement(v, s / 2 - 1)];
T b = v[kthElement(v, s / 2)];
return (a + b) / 2;
}
#include < cstdio >
#include < algorithm >
int main()
{
int n;
while(scanf("%d", &n) == 1)
{
int t;
vector<int> v;
for (int i = 0; i < n; i++)
{
// scanf("%d", &t);
v.push_back(i);
//v.push_back(t);
}
random_shuffle(v.begin(), v.end());
//printf("%d\n", v[kthElement(v, n / 2)]);
printf("%d\n", median(v));
}
return 0;
}
#include < set >
using namespace std;
#include < cstdlib >
#include < cassert >
size_t rand(size_t maxn)
{
size_t d = RAND_MAX / maxn;
size_t nd = d * maxn;
size_t r;
do r = rand(); while(r >= nd);
return r % maxn;
}
int rand( int a, int b)
{
assert(a < b);
return (int)rand(b - a) + a;
}
template < typename T >
size_t kthElement( const vector < T > & v, size_t k)
{
vector<size_t> s(v.size());
for (size_t i = 0; i < v.size(); i++)
s[i] = i;
size_t curK = k;
size_t t;
for(;;)
{
size_t c = 0;
t = rand(s.size());
const T &value = v[s[t]];
for (vector<size_t>::const_iterator it = s.begin(); it != s.end(); ++it)
{
if (v[*it] < value)
c++;
}
if (c == curK)
break;
swap(s[t], s.back());
s.pop_back();
// now, the v[s[t]] is c-th element of the rest v
if (c > curK)
{
// target is smaller than v[s[t]]
// remove greater values from s
// s.erase(remove_if(s.begin(), s.end(), bind2st(greater<T>(), value)), s.end());
for (vector<size_t>::iterator it = s.begin(); it != s.end();)
{
if (v[*it] > value)
swap(*it, s.back()), s.pop_back();
else
++it;
}
}
else if (c < curK)
{
// target is greater than v[s[t]]
// remove smaller values from s
for (vector<size_t>::iterator it = s.begin(); it != s.end();)
{
if (v[*it] < value)
swap(*it, s.back()), s.pop_back();
else
++it;
}
curK -= c + 1;
}
else
break;
}
return s[t];
}
template < typename T >
T median( const vector < T > & v)
{
size_t s = v.size();
if (s & 1)
{
return v[kthElement(v, s / 2)];
}
T a = v[kthElement(v, s / 2 - 1)];
T b = v[kthElement(v, s / 2)];
return (a + b) / 2;
}
#include < cstdio >
#include < algorithm >
int main()
{
int n;
while(scanf("%d", &n) == 1)
{
int t;
vector<int> v;
for (int i = 0; i < n; i++)
{
// scanf("%d", &t);
v.push_back(i);
//v.push_back(t);
}
random_shuffle(v.begin(), v.end());
//printf("%d\n", v[kthElement(v, n / 2)]);
printf("%d\n", median(v));
}
return 0;
}