Description

A young schoolboy would like to calculate the sum


for some fixed natural k and different natural n. He observed that calculating ik
for all i (1<=i<=n) and summing up results is a too slow way to
do it, because the number of required arithmetical operations increases
as n increases. Fortunately, there is another method which takes only a
constant number of operations regardless of n. It is possible to show
that the sum Sk(n) is equal to some polynomial of degree k+1 in the variable n with rational coefficients, i.e.,



We require that integer M be positive and as small as possible. Under this condition the entire set of such numbers (i.e. M, ak+1, ak, ... , a1, a0)
will be unique for the given k. You have to write a program to find
such set of coefficients to help the schoolboy make his calculations
quicker.

Input

The input file contains a single integer k (1<=k<=20).

Output

Write integer numbers M, ak+1, ak, ... , a1, a0
to the output file in the given order. Numbers should be separated by
one space. Remember that you should write the answer with the smallest
positive M possible.

Sample Input

2

Sample Output

6 2 3 1 0

Source

【分析】
题意就是给出一个k,找一个最小的M使得中a[i]皆为整数.
这个涉及到伯努利数的一些公式,如果不知道的话基本没法做..

1. 伯努利数与自然数幂的关系:

2. 伯努利数递推式:

先通过递推式求得伯努利数,然后用1公式并将中间的(n+1) ^ i,变成n ^ i,后面再加上n ^ k,化进去就行了。

 /*
宋代朱敦儒
《西江月·世事短如春梦》
世事短如春梦,人情薄似秋云。不须计较苦劳心。万事原来有命。
幸遇三杯酒好,况逢一朵花新。片时欢笑且相亲。明日阴晴未定。
*/
#include <cstdio>
#include <cstring>
#include <algorithm>
#include <cmath>
#include <queue>
#include <vector>
#include <iostream>
#include <string>
#include <ctime>
#define LOCAL
const int MAXN = + ;
const double Pi = acos(-1.0);
using namespace std;
typedef long long ll;
ll gcd(ll a, ll b){return b == ? a: gcd(b, a % b);}
struct Num{
ll a, b;//分数,b为分母
Num(ll x = , ll y = ) {a = x;b = y;}
void update(){
ll tmp = gcd(a, b);
a /= tmp;
b /= tmp;
}
Num operator + (const Num &c){
ll fz = a * c.b + b * c.a, fm = b * c.b;
if (fz == ) return Num(, );
ll tmp = gcd(fz, fm);
return Num(fz / tmp, fm / tmp);
}
}B[MAXN], A[MAXN];
ll C[MAXN][MAXN]; void init(){
//预处理组合数
for (int i = ; i < MAXN; i++) C[i][] = C[i][i] = ;
for (int i = ; i < MAXN; i++)
for (int j = ; j < MAXN; j++) C[i][j] = C[i - ][j] + C[i - ][j - ];
//预处理伯努利数
B[] = Num(, );
for (int i = ; i < MAXN; i++){
Num tmp = Num(, ), add;
for (int j = ; j < i; j++){
add = B[j];
add.a *= C[i + ][j];
tmp = tmp + add;
}
if (tmp.a) tmp.b *= -(i + );
tmp.update();
B[i] = tmp;
}
}
void work(){
int n;
scanf("%d", &n);
ll M = n + , flag = , Lcm;
A[] = Num(, );
for (int i = ; i <= n + ; i++){
if (B[n + - i].a == ) {A[i] = Num(, );continue;}
Num tmp = B[n + - i];
tmp.a *= C[n + ][i];//C[n+1][i] = C[n + 1][n + 1 - i]
tmp.update();
if (flag == ) Lcm = flag = tmp.b;
A[i] = tmp;
}
A[n] = A[n] + Num(n + , ); for (int i = ; i <= n + ; i++){
if (A[i].a == ) continue;
Lcm = (Lcm * A[i].b) / gcd(Lcm, A[i].b);
}
if (Lcm < ) Lcm *= -;
M *= Lcm;
printf("%lld", M);
for (int i = n + ; i >= ; i--) printf(" %lld", A[i].a * Lcm / A[i].b);
} int main(){ init();
work();
//printf("%lld\n", C[5][3]);
return ;
}

【POJ1707】【伯努利数】Sum of powers的更多相关文章

  1. [伯努利数] poj 1707 Sum of powers

    题目链接: http://poj.org/problem?id=1707 Language: Default Sum of powers Time Limit: 1000MS   Memory Lim ...

  2. [CSAcademy]Sum of Powers

    [CSAcademy]Sum of Powers 题目大意: 给定\(n,m,k(n,m,k\le4096)\).一个无序可重集\(A\)为合法的,当且仅当\(|A|=m\)且\(\sum A_i=n ...

  3. Euler's Sum of Powers Conjecture

    转帖:Euler's Sum of Powers Conjecture 存不存在四个大于1的整数的五次幂恰好是另一个整数的五次幂? 暴搜:O(n^4) 用dictionary:O(n^3) impor ...

  4. UVA766 Sum of powers(1到n的自然数幂和 伯努利数)

    自然数幂和: (1) 伯努利数的递推式: B0 = 1 (要满足(1)式,求出Bn后将B1改为1 /2) 参考:https://en.wikipedia.org/wiki/Bernoulli_numb ...

  5. UVa 766 Sum of powers (伯努利数)

    题意: 求 ,要求M尽量小. 析:这其实就是一个伯努利数,伯努利数公式如下: 伯努利数满足条件B0 = 1,并且 也有 几乎就是本题,然后只要把 n 换成 n-1,然后后面就一样了,然后最后再加上一个 ...

  6. POJ 1707 Sum of powers(伯努利数)

    题目链接:http://poj.org/problem?id=1707 题意:给出n 在M为正整数且尽量小的前提下,使得n的系数均为整数. 思路: i64 Gcd(i64 x,i64 y) { if( ...

  7. sum of powers

    题意: 考虑所有的可重集{a1,a2,a3....ak} 满足a1+a2+....+ak=n,求所有a1^m+a2^m+a3^m的和 n,m,k<=5000 题解: part1: 考虑f[i][ ...

  8. 51nod1228 序列求和(自然数幂和)

    与UVA766 Sum of powers类似,见http://www.cnblogs.com/IMGavin/p/5948824.html 由于结果对MOD取模,使用逆元 #include<c ...

  9. [转] Loren on the Art of MATLAB

    http://blogs.mathworks.com/loren/2007/03/01/creating-sparse-finite-element-matrices-in-matlab/ Loren ...

随机推荐

  1. Bzoj 3289: Mato的文件管理 莫队,树状数组,逆序对,离散化,分块

    3289: Mato的文件管理 Time Limit: 40 Sec  Memory Limit: 128 MBSubmit: 1539  Solved: 665[Submit][Status][Di ...

  2. Preferred Java way to ping a HTTP Url for availability

    I need a monitor class that regularly checks whether a given HTTP URL is available. I can take care ...

  3. windows mobile 6.5 隐藏 左下角(左上角)的开始按钮 叉号关闭按钮

    其实做起来很简单,但是国内的网站就是找不到. 1.开始按钮原来的界面是这样的: windows mobile 6.0界面: windows mobile 6.5.X界面: 修改一个windows mo ...

  4. spfa + slf优化

    最近在练习费用流 , 不是要用spfa吗 ,我们教练说:ns学生写朴素的spfa说出去都让人笑 . QwQ,所以就去学了一下优化 . slf优化就是双向队列优化一下,本来想用lll优化,可是优化后我t ...

  5. Oracle DB 自动管理共享内存

    • 启用Oracle Enterprise Manager (EM) 内存参数 • 设置自动优化的内存参数 • 使用手动优化的SGA 参数覆盖最小大小 • 使用SGA Advisor 设置SGA_TA ...

  6. pgsql自动安装shell脚本整理

    前面不断在vm虚拟机上测试pgsql,发觉安装还是有些麻烦的. 所以就收集了一些 1,http://www.davidghedini.com/pg/entry/postgresql_9_5_scrip ...

  7. MP算法和OMP算法及其思想

    主要介绍MP(Matching Pursuits)算法和OMP(Orthogonal Matching Pursuit)算法[1],这两个算法尽管在90年代初就提出来了,但作为经典的算法,国内文献(可 ...

  8. Android将Activity打成jar包供第三方调用(解决资源文件不能打包的问题)

    转载地址:http://blog.csdn.net/xiaanming/article/details/9257853 最近有一个需要,我们公司做了一个apk客户端,然后其他的公司可以根据自己的需要来 ...

  9. Linux内核源代码情景分析系列

    http://blog.sina.com.cn/s/blog_6b94d5680101vfqv.html Linux内核源代码情景分析---第五章 文件系统  5.1 概述 构成一个操作系统最重要的就 ...

  10. getline和get的区别

    #include<iostream> #include<fstream> #include<cstring> using namespace std; int ma ...