hdu5608 几类经典的根号复杂度算法

I, as a ACMer, always take algorithm complexity into consideration when programming. Today, I introduce you some elegant algorithms of root Complex.

转载请注明Reproduced please specifydna049.com

1. $s(n) = \sum_{i=1}^{n} \lfloor \frac{n}{i} \rfloor$

Since The range of $\lfloor \frac{n}{i} \rfloor$ contains at most $2\sqrt{n}$ . There may exist a algorithm of complexity $O(\sqrt{n})$.

LL getsum(LL n){ // The code is simple and easy to understand
    LL sum = 0;
    for(LL i=1,j;i<=n;i=j+1){ 
        j = n/(n/i);
        sum += (j-i+1)*(n/i);
    }
    return sum;
}

Actually, $s(n)$ donate the number of positive integer point under graph $xy=1$.

2. $\sigma_k(n) = \sum_{d|n} d^k$

1. $\sigma_0(n)$ donate the number of divisors.
2. $\sigma_1(n)$ donate the sum of divisors.

LL mypow(LL x,LL n){
    LL r = 1;
    while(n){
        if(n&1) r=r*x;
        n>>=1;  x=x*x;
    }
    return r;
}
LL getr(LL n,LL k){
    LL r = 0,d;
    for(d=1;d*d<n;++d){
        if(n%d==0)  r += mypow(d,k) + mypow(n/d,k);
    }
    if(d*d == n) r+=mypow(d,k);
    return r;
}

The code above is primary and trival.

3. $f(n) = \sum_{i=1}^n \sigma_k(i)$

f(n)=∑i=1nσk(n)=∑i=1