[PAT] Perfect Sequence

题目

Given a sequence of positive integers and another positive integer p. The sequence is said to be a "perfect sequence" if M <= m * p where M and m are the maximum and minimum numbers in the sequence, respectively.

Now given a sequence and a parameter p, you are supposed to find from the sequence as many numbers as possible to form a perfect subsequence.

Input Specification:

Each input file contains one test case. For each case, the first line contains two positive integers N and p, where N (<= 10^5^) is the number of integers in the sequence, and p (<= 10^9^) is the parameter. In the second line there are N positive integers, each is no greater than 10^9^.

Output Specification:

For each test case, print in one line the maximum number of integers that can be chosen to form a perfect subsequence.

Sample Input:

10 8
2 3 20 4 5 1 6 7 8 9

Sample Output:

8

给一个数字数组，给一个系数p，在数组中取尽量多的数组成集合A，使得A中最大值max和最小值min满足max <= min*p。返回A的规模。

分析

很简单，先把数组从小到大排序，用left、right两个指针扫描，left从0到n-1停止，right在left的循环内层，每次从left到n-1或遇到right > left * p停止。用tmp记录从当前left开始的perfect sequence的长度，maxSize记录最终结果。

特别地，如果某一次扫描中right到达了数组的左边界，此时的tmp一定是最大长度，直接输出就可以了。

要注意的是，p最大9位，2^31是10位，用int型的话，乘积溢出的风险很大，所以p要用long long型。

代码

#include <iostream>
#include <vector>
#include <algorithm>
using namespace std;

int main() {
	int n;
	long long p;
	scanf("%d%lld", &n, &p);
	vector<int> arr(n);
	for (int i = 0; i < n; i++) {
		cin >> arr[i];
	}
	sort(arr.begin(), arr.end());
	int maxSize = 0, tmp = 0;
	for (int left = 0; left < n; left++) {
		for (int right = left + maxSize; right < n; right++) {
			if (arr[right] <= arr[left] * p) {
				tmp = right - left + 1;
				if (right == n - 1) {
					printf("%d", tmp);
					return 0;
				}
				maxSize = tmp > maxSize ? tmp: maxSize;
			}
			else {
				break;
			}
		 }
	}
	printf("%d", maxSize);
	return 0;
}

复杂度

代码写了两层循环，看起来像是个O(n^2)的算法。实际上，每次left移动后。right的扫描都是从left + maxSize开始的，而并非从left开始。也就是说，不论left移动还是right移动，right每次都要向右移动一步，主算法复杂度O(n)。

真正影响效率的是排序的O(nlogn)。