題解/算法 {196. 质数距离}

「已注销」

已于 2023-11-13 15:31:41 修改

阅读量65

点赞数

分类专栏：题解文章标签：算法

于 2023-01-13 17:16:12 首次发布

本文链接：https://blog.csdn.net/qq_66485519/article/details/128677128

版权

题解专栏收录该内容

298 篇文章 0 订阅

订阅专栏

文章讨论了一个在大范围内（不超过2^9）查找素数的问题，利用关键性质：如果一个数不是素数，那么它至少有一个质因数不大于其平方根。因此，可以只关注小于等于范围平方根的质数，筛去它们的所有倍数，从而找出素数。然而，这种方法不会筛掉0和1，需要额外处理。文章提供了一种算法实现，并指出了可能的错误情况，如计算范围的处理和倍数检查的条件等。

摘要由CSDN通过智能技术生成

Solution

Link

Solution

The problem is seemingly unsolvable (iterating all prime-numbers between $[L, R]$ where $\leq L \leq R \leq 2e9$ ), however, it’s vital to real the problem-content carefully, there maybe a hint ( $\leq 1e6$ ) in a corner;

--

There is an important property (a common and useful Trick) for an integer $x$ :
. If $x$ is not a Prime-Number, then $min(p_i) \leq \sqrt{ x}$ (in other words, there must exist a Prime-Factors $p$ of $x$ such that $\ \ \land \ \ p \in [1, \sqrt{x}]$ )

As a result, we can stand in the converse view-point, focusing on all Primes $p = [1, 1 e 6]$ instead of $x$ , then any Multiple of $p$ must be not a Prime-Number;
. Note that, this method would only sieved all Non-Prime-Numbers $a$ which contains a Prime-Number, that is, $a = 0/1$ would not be sieved although they are also Non-Primes;

Finding out these Prime-Numbers in the range $[l, r]$ where $\leq l \leq r \leq 1e12 \quad \land (r-l) \leq 1e6$ ; the algorithm is:

int l, r;
bool Is_prime[ r - l + 1] = true;
for( p : $(all Primes `<= \sqrt{r}`){ //< any Non-Prime in [l,r] must has a Prime which is `<= \sqrt{r}`;
	for( a : [$(all number in [l,r])]_AND_[$([k*p]_AND_[k>=2] (the multiple of `p`))]){ //< @Location-1
	//< `a` is a multiple of `p`, and also `a != p` (very important); 
		Is_prime[ a - l] = false;
	}
}
//>< note that, only a Non-Prime which has a Prime-Factor (>=2) would be sieved (set `false`); in other words, although `0,1` are also Non-Primes, they would not be sieved (you need check them specially); 
vector<> primes;
for( i : [l,r]){
	if( (i >= 2) && Is_prime[ i]){ //< @Location-0
		primes.push_back( i);
	}
}

But when you applied, it is very fallible, let’s see some mistakes:
+ If int r, then for( int i = l; i <= r; ++i) would be an error, cuz r would equals INT_MAX;
+ @Location-0, the additional i >= 2;
+ @Location-1, $a$ is not the multiple of $p$ in the range $[l, r]$ , you must make sure $a > p$ ;