HDU BestCoder 1001

BestCoder Sequence

Time Limit: 2000/1000 MS (Java/Others)    Memory Limit: 32768/32768 K (Java/Others)
Total Submission(s): 0    Accepted Submission(s): 0

Problem Description

  
  

   
   Mr Potato is a coder.
Mr Potato is the BestCoder.

One night, an amazing sequence appeared in his dream. Length of this sequence is odd, the median number is M, and he named this sequence as 
   
   Bestcoder Sequence.

As the best coder, Mr potato has strong curiosity, he wonder the number of consecutive sub-sequences which are 
   
   bestcoder sequences in a given permutation of 1 ~ N.

  
  

 

Input

  
  

   
   Input contains multiple test cases. 
For each test case, there is a pair of integers N and M in the first line, and an permutation of 1 ~ N in the second line.


   
   [Technical Specification]
1. 1 <= N <= 40000
2. 1 <= M <= N

  
  

 

Output

  
  

   
   For each case, you should output the number of consecutive sub-sequences which are the 
   
   Bestcoder Sequences. 

  
  

 

Sample Input

  
  

   
   1 1
1
5 3
4 5 3 2 1
  
  

 

Sample Output

  
  

   
   1
3


   
   

    
    

     
     Hint
    
    
For the second case, {3},{5,3,2},{4,5,3,2,1} are Bestcoder Sequence.

   
   
   
   


   
   
   
   

    
    下面是我的提交代码，改了三次，用了我觉得最快的算法了，但还是超时了
   
   
   
   


   
   
   
   

    
    
#include <stdio.h>
#define loop(i, n) for(int i=0; i<n; ++i)

int p[40000];
int b[40000];
int s[40000];

int main()
{
	int n, m, index, ans;
	while(1)
	{
		if(scanf("%d%d",&n,&m) != 2) break;
		loop(i, n)
		{
			scanf("%d",&p[i]);
			if(p[i] == m) index = i;
		}
		
		b[index] = s[index] = 0;
		for(int i=index-1; i>=0; --i)
		{
			if(p[i] > m) 
			{
				b[i] = b[i+1] + 1;
				s[i] = s[i+1];
			}
			else
			{
				s[i] = s[i+1] + 1;
				b[i] = b[i+1];
			}
		}
		for(int i=index+1; i<n; ++i)
		{
			if(p[i] > m)
			{
				b[i] = b[i-1] + 1;
				s[i] = s[i-1];
			}
			else
			{
				s[i] = s[i-1] + 1;
				b[i] = b[i-1];
			}
		}
		ans = 0;
		for(int beg=index; beg>=0; --beg)
		{
			for(int end=index; end<n; ++end)
			{
				if((end-beg)%2==1) continue;
				else if(b[beg]+b[end] == s[beg]+s[end]) ++ans;
			}
		}
		printf("%d\n",ans);
	}
	return 0;
}
    
    

    
    

   
   
   
   

    
    无语死，待会看一下别人是怎么个快法吧！！！
   
   
   
   

    
    我自己是从M的两端扩展，记录比M大和小的个数，然后再枚举所有包含M的子序列，判断是否符合M是序列的中位数的条件（就是M两边比它大的总数是否和M两边比它小的总数相等），如果是的话，就++ans；
   
   
   
   

    
    这样最后就算出来啦~然后就超时啦！
   
   
   
   

    
    哈哈啊~~痛呀，不过第一道还是1A了，挺开心的，因为上一轮，我连一道都没过，这次进步了！加油！