Mike has a sequence A = [a1, a2, ..., an] of length n. He considers the sequence B = [b1, b2, ..., bn] beautiful if the gcd of all its elements is bigger than 1, i.e. .

Mike wants to change his sequence in order to make it beautiful. In one move he can choose an index i (1 ≤ i < n), delete numbers ai, ai + 1 and put numbers ai - ai + 1, ai + ai + 1 in their place instead, in this order. He wants perform as few operations as possible. Find the minimal number of operations to make sequence A beautiful if it's possible, or tell him that it is impossible to do so.

is the biggest non-negative number d such that d divides bi for every i (1 ≤ i ≤ n).

Input

The first line contains a single integer n (2 ≤ n ≤ 100 000) — length of sequence A.

The second line contains n space-separated integers a1, a2, ..., an (1 ≤ ai ≤ 109) — elements of sequence A.

Output

Output on the first line "YES" (without quotes) if it is possible to make sequence A beautiful by performing operations described above, and "NO" (without quotes) otherwise.

If the answer was "YES", output the minimal number of moves needed to make sequence A beautiful.

Example

Input
2
1 1
Output
YES
1
Input
3
6 2 4
Output
YES
0
Input
2
1 3
Output
YES
1

Note

In the first example you can simply make one move to obtain sequence [0, 2] with .

In the second example the gcd of the sequence is already greater than 1.

题意:n个数,n<=1e5,操作:把a[i],a[i+1] 替换成 a[i]-a[i+1],a[i]+a[i+1],问至少要多少次操作才能让整个a数组的最大公约数gcd大于1.

由题目给出操作可知:当gcd(a,b)<=1时,进行操作为:

初始:a  b

第一步:a-b  a+b

第二步:-2b  2a

即两个数最多2步操作就能满足GCD==2。

对于两个偶数,要进行0步操作;对于两个奇数,要进行1步操作;对于一个奇数一个偶数,要进行2步操作。

先把所有“2个奇数成对”的情况计数+1并把两个奇数更新为偶数,然后在重新判断所有“1个奇数1个偶数成对”的情况计数+2。

#include<bits/stdc++.h>
using namespace std;
typedef long long ll;
ll a[200050],n,num=0;
ll gcd(ll a,ll b){
return b==0?a:gcd(b,a%b);
}
int main(){
cin>>n;
for(int i=1;i<=n;i++)
scanf("%lld",&a[i]);
ll ans=gcd(abs(a[1]),abs(a[2]));
for(int i=3;i<=n;i++)\
ans=gcd(ans,abs(a[i]));
if(ans>1) cout<<"YES"<<endl<<0<<endl;
else
{
for(int i=1;i<n;i++)
if(a[i]%2&&a[i+1]%2)
a[i]=0,a[i+1]=0,num++;
for(int i=1;i<n;i++)
if((a[i]%2&&a[i+1]%2==0)||(a[i]%2==0&&a[i]%2))
a[i]=0,a[i+1]=0,num+=2;
cout<<"YES"<<endl<<num<<endl;
}
return 0;
}

CodeForce-798C Mike and gcd problem(贪心)的更多相关文章

  1. Codeforces 798C - Mike and gcd problem(贪心+数论)

    题目链接:http://codeforces.com/problemset/problem/798/C 题意:给你n个数,a1,a2,....an.要使得gcd(a1,a2,....an)>1, ...

  2. Codeforces 798C. Mike and gcd problem 模拟构造 数组gcd大于1

    C. Mike and gcd problem time limit per test: 2 seconds memory limit per test: 256 megabytes input: s ...

  3. codeforces 798c Mike And Gcd Problem

    题意: 给出一个数列,现在有一种操作,可以任何一个a[i],用a[i] – a[i+1]和a[i]+a[i+1]替代a[i]和a[i+1]. 问现在需要最少多少次操作,使得整个数列的gcd大于1. 思 ...

  4. codeforces 798C.Mike and gcd problem 解题报告

    题目意思:给出一个n个数的序列:a1,a2,...,an (n的范围[2,100000],ax的范围[1,1e9] ) 现在需要对序列a进行若干变换,来构造一个beautiful的序列: b1,b2, ...

  5. CF798 C. Mike and gcd problem

    /* CF798 C. Mike and gcd problem http://codeforces.com/contest/798/problem/C 数论 贪心 题意:如果一个数列的gcd值大于1 ...

  6. 【算法系列学习】codeforces C. Mike and gcd problem

    C. Mike and gcd problem http://www.cnblogs.com/BBBob/p/6746721.html #include<iostream> #includ ...

  7. Codeforces Round #410 (Div. 2)C. Mike and gcd problem

    题目连接:http://codeforces.com/contest/798/problem/C C. Mike and gcd problem time limit per test 2 secon ...

  8. codeforces#410C Mike and gcd problem

    题目:Mike and gcd problem 题意:给一个序列a1到an ,如果gcd(a1,a2,...an)≠1,给一种操作,可以使ai和ai+1分别变为(ai+ai+1)和(ai-ai+1); ...

  9. #410div2C. Mike and gcd problem

    C. Mike and gcd problem time limit per test 2 seconds memory limit per test 256 megabytes input stan ...

  10. Mike and gcd problem CodeForces - 798C (贪心思维+数论)

    题目链接 比较棒的一道题, 题意: 给你一个N个数的数组,让你用尽量少的操作使整个数组的gcd大于1,即gcd(a1 ,a2,,,,an) > 1 如果可以输出YES和最小的次数,否则输出NO ...

随机推荐

  1. 【javaFX学习】(二) 面板手册

    移至http://blog.csdn.net/qq_37837828/article/details/78732591 更新 找了好几个资料,没找到自己想要的,自己整理下吧,方便以后用的时候挑选,边学 ...

  2. IP地址,InetAddress类的使用

    IP地址 IP地址:InetAddress(没有构造器,通过静态方法返回) java.net包下 唯一定位一台网络上的计算机 127.0.0.1:本机localhost ip地址的分类 IPV4/IP ...

  3. putty编译过程

    在Win7上用Visual Studio编译putty源代码. 安装vs2005,只安装c++和.net framework sdk即可: 将putty-src.zip解压到e:\MyDoc\VSPr ...

  4. Linux平台上转换文件编码

    Linux系统的iconv指令是一个很好的文件编码转换工具,支持的编码范围广,使用方便,例如将一个utf-8编码的文件(名为tic)转换为gbk编码: iconv -f utf-8 -t gbk ti ...

  5. 服务启动shell脚本

    #!/bin/sh JarDir=`pwd` do_start() { echo "pandora-login start ..." nohup java -jar -Xmn256 ...

  6. 内网隧道与SOCKS代理思路总结

    早就想总结一下这部分内容了,总是有其他事情卡住,一直拖到现在 内网中有很多边界设备,比如防火墙.这种边界设备会控制内部主机的对外连接,一般会仅允许某些种类的端口开放或某种数据流量出入 这就意味着我们只 ...

  7. NOIP 模拟 $16\; \rm Lost My Music$

    题解 \(by\;zj\varphi\) 一道凸包的题 设 \(\rm dep_u\) 表示节点 \(u\) 的深度,那么原式就可化为 \(-\frac{c_v-c_u}{dep_v-dep_u}\) ...

  8. 题解 graph

    传送门 一道做了巨久,不过确实很好的题 发现不定边权极难处理,所以就不会 感觉和这题有点像,但还是不会 但发现题面里有个地方很套路 要求有哪些点/边最终可以满足最短/最小,比如这样或这样的题,考虑凸包 ...

  9. 算法入门 - 基于动态数组的栈和队列(Java版本)

    之前我们学习了动态数组的实现,接下来我们用它来实现两种数据结构--栈和队列.首先,我们先来看一下栈. 什么是栈? 栈是计算机的一种数据结构,它可以临时存储数据.那么它跟数组有何区别呢? 我们知道,在数 ...

  10. ffmpeg第6篇:滤镜语法

    前言 哈哈,回来继续填坑了,前段时间较忙没时间写,现在继续~ 简介 滤镜是ffmpeg的一个很强大的功能,它支持许多有用的视频处理功能,常见的滤镜如:缩放.旋转.水印.裁剪等 一个比较经典的滤镜使用方 ...