Educational Codeforces Round 22 补题 CF 813 A-F
直接粗暴贪心 略过
#include<bits/stdc++.h>
using namespace std;
int main()
{//freopen("t.txt","r",stdin);
long long int n,sum=0,k;
scanf("%lld",&n);
for(int i=0;i<n;i++)
{
scanf("%lld",&k);
sum+=k;
}
scanf("%lld",&k);
long long int l,r;
for(int i=0;i<k;i++)
{
scanf("%lld%lld",&l,&r);
if(sum<=r){printf("%lld\n",max(sum,l));return 0;}
}
printf("-1\n");
return 0;
}
由于指数增长速度极快 直接用一个平衡树遍历即可。
#include<bits/stdc++.h>
using namespace std;
typedef long long int LL;
const LL N=1e18+1;
LL x[100],y[100];
map<LL,LL>s;
int main()
{//freopen("t.txt","r",stdin);
LL l,r;
s.clear();
scanf("%lld%lld%lld%lld",&x[1],&y[1],&l,&r);
x[0]=y[0]=1;
for(int i=2;i<100&&x[i-1]*x[1]<=N&&x[i-1]<=(N/x[1])&&x[i-1]*x[1]>=0&&(!(x[i-1]>1e+9&&x[1]>1e+9));i++)
x[i]=x[i-1]*x[1];
for(int i=2;i<100&&y[i-1]*y[1]<=N&&y[i-1]<=(N/y[1])&&y[i-1]*y[1]>=0&&(!(y[i-1]>1e+9&&y[1]>1e+9));i++)
y[i]=y[i-1]*y[1];
for(int a=0;x[a]<=N&&x[a]!=0;a++)
for(int b=0;y[b]<=N&&y[b]!=0;b++ )
{
//if(a==b)continue;
if(x[a]<0||y[b]<0)continue;
if(x[a]+y[b]<=r&&x[a]+y[b]>=l&&s.count(x[a]+y[b])==0)s[x[a]+y[b]]=1;
}
map<LL,LL>::iterator it,ipt;
LL ans=0,len=0;
s[l-1]=s[r+1]=1;
it=s.begin();
ipt=s.begin();
it++;
for(;it!=s.end();it++,ipt++)
{
//cout<<it->first<<" "<<ipt->first<<endl;
ans=max(ans,it->first-ipt->first-1);
}
printf("%lld\n",ans);
return 0;
}
很简单的贪心 Bob总是要前往最深的节点(在不撞到Alice的前提下)
遍历一下即可
#include<bits/stdc++.h>
using namespace std;
const int N=2e5+1;
int fa[N],dep[N],len[N];
vector<int>adj[N];
int n,x;
int dfs(int cur,int father,int lens)
{
dep[cur]=lens;
len[cur]=0;
fa[cur]=father;
for(int i=0;i<adj[cur].size();i++)
{
int ne=adj[cur][i];
if(ne==father)continue; len[cur]=max(len[cur],dfs(ne,cur,lens+1));
}
return len[cur]+1;
}
int main()
{//freopen("t.txt","r",stdin);
scanf("%d%d",&n,&x);
int a,b;
for(int i=0;i<n;i++)
{
scanf("%d%d",&a,&b);
adj[a].push_back(b);
adj[b].push_back(a);
}
fa[1]=0;
dfs(1,0,0);
int ll=0;
int ans=dep[x]+len[x];
while((x=fa[x])!=0)
{
ll++;
if(dep[x]<=ll)break;
ans=max(ans,dep[x]+len[x]);
}
printf("%d\n",ans*2);
return 0;
}
O(n^2)的dp 设dp[i][j]为 第一个Melody在i处结束 第二个Melody在j处结束 显然有 dp[i][j]=dp[j][i]
具体转移方法看代码吧 很清晰
# include <iostream>
#include<bits/stdc++.h>
using namespace std;
long long i,j,n,mx,a[5009],b[5009][5009],x[9],y[100009];
int main()
{
cin>>n;
for (i=1;i<=n;i++) cin>>a[i];
for (i=0;i<=n;i++)
{
for (j=0;j<=7;j++) x[j]=0;
for (j=1;j<=n;j++) y[a[j]]=0;
for (j=1;j<i;j++)
{
x[a[j]%7]=max(x[a[j]%7],b[i][j]);
y[a[j]]=max(y[a[j]],b[i][j]);
}
for (j=i+1;j<=n;j++)
{
b[i][j]=max(max(y[a[j]+1],y[a[j]-1]),max(x[a[j]%7],b[i][0]))+1;
b[j][i]=b[i][j];
x[a[j]%7]=max(x[a[j]%7],b[i][j]);
y[a[j]]=max(y[a[j]],b[i][j]);
mx=max(mx,b[i][j]);
}
}
cout<<mx;
}
技巧性很强的一道线段树题
我们首先考虑解决一个问题:对于某个士兵,它有可能对哪个区间造成超员?
假设该士兵的位置是d 对于同种类的士兵 往前数k个的位置的士兵 它的位置是b 那么对于区间【1,b】 位置d的士兵会造成他们超员
用线段树可以在对数时间内解决第一个问题。
那么对于题中询问[L,R] 我们只要统计使得区间[L,R]可能超员的士兵 位置小于等于R的个数即可。
总的复杂度 O(nlog^2n)
不过这个代码写的和一般的线段树还不太一样
#include<bits/stdc++.h>
using namespace std; const int Maxn = 100005;
const int Maxm = 524288; int n, k;
int a[Maxn];
vector <int> I[Maxn];
vector <int> st[Maxm];
int q;
int res; void Insert(int v, int l, int r, int a, int b, int val)
{
if (l == a && r == b) st[v].push_back(val);
else {
int m = l + r >> 1;
if (a <= m) Insert(2 * v, l, m, a, min(m, b), val);//重复添加
if (m + 1 <= b) Insert(2 * v + 1, m + 1, r, max(m + 1, a), b, val);
}
} int Get(int v, int l, int r, int x, int R)
{
int res = upper_bound(st[v].begin(), st[v].end(), R) - st[v].begin();
if (l < r) {
int m = l + r >> 1;
if (x <= m) res += Get(2 * v, l, m, x, R);//避免重复计算
else res += Get(2 * v + 1, m + 1, r, x, R);
}
return res;
} int main()
{
scanf("%d %d", &n, &k);
for (int i = 1; i <= n; i++) {
scanf("%d", &a[i]);
I[a[i]].push_back(i);
if (I[a[i]].size() > k) Insert(1, 1, n, 1, I[a[i]][I[a[i]].size() - k - 1], i);
}
scanf("%d", &q);
while (q--) {
int x, y; scanf("%d %d", &x, &y);
x = (x + res) % n + 1;
y = (y + res) % n + 1;
if (x > y) swap(x, y);
res = (y - x + 1) - Get(1, 1, n, x, y);
printf("%d\n", res);
}
return 0;
}
很具有扩展性的一道题 可以联想到很多问题
官方题解很详细 直接引用
简单概述一下:我们转化问题 对于每条边 我们考虑它覆盖了那些询问 比如对于第i个询问 我们只需要考虑覆盖了它的边即可。
用分治法dfs的时候 我们需要添加边,可以简单做到,删除边该怎么做呢? 把图的信息当作一个栈回溯即可。(分治的过程满足后进先出的性质,所以可以用栈)
复杂度不高,但是常数真的大。
下面引用原文。
If the edges were only added and not deleted, it would be a common problem that is solved with disjoint set union. All you need to do in that problem is implement a DSU which maintains not only the leader in the class of some vertex, but also the distance to this leader. Then, if we try to connect two vertices that have the same leader in DSU and the sum of distances to this leader is even, then we get a cycle with odd length, and graph is no longer bipartite.
But in this problem we need to somehow process removing edges from the graph.
In the algorithm I will describe below we will need to somehow remove the last added edge from DSU (or even some number of last added edges). How can we process that? Each time we change some variable in DSU, we can store an address of this variable and its previous value somewhere (for example, in a stack). Then to remove last added edge, we rollback these changes — we rewrite the previous values of the variables we changed by adding the last edge.
Now we can add a new edge and remove last added edge. All these operations cost (O(logn)) because we won't use path compression in DSU — path compression doesn't work in intended time if we have to rollback. Let's actually start solving the problem.
For convinience, we change all information to queries like "edge (x, y) exists from query number l till query number r". It's obvious that there are no more than q such queries. Let's use divide-and-conquer technique to make a function that answers whether the graph is bipartite or not after every query from some segment of queries [a, b]. First of all, we add to DSU all the edges that are present in the whole segment (and not added yet); then we solve it recursively for
and
; then we remove edges from DSU using the rollback technique described above. When we arrive to some segment [a, a], then after adding the edges present in this segment we can answer if the graph is bipartite after query a. Remember to get rid of the edges that are already added and the edges that are not present at all in the segment when you make a recursive call. Of course, to solve the whole problem, we need to call our function from segment [1, q].
Time complexity is
, because every edge will be added only in
calls of the function.
#include<iostream>
#include<cstdio>
#include<map>
#define N 500005
using namespace std;
int n,m,T,fa[N],top,st[N<<2],a[N];
struct E{int x,y,l,r;}e[N];
int Getfa(int x)
{
while (fa[x]!=x) x=fa[x];
return x;
}
int col(int x)//??????
{
int now=0;
while (fa[x]!=x) now^=a[x],x=fa[x];//???????????????
return now;
}
void link(int x,int y,int D)
{ fa[x]=y;a[x]=D;st[++top]=x;
}
void RE(int t)
{
for (;top>t;top--)
fa[st[top]]=st[top],a[st[top]]=0;
}
void work(int l,int r,int k)
{
// cout<<l<<r<<k<<endl;
int mid=(l+r)>>1,now=top;
for (int i=1;i<=k;i++)
if (e[i].l<=l&&r<=e[i].r)//????????????
{ int u=Getfa(e[i].x),v=Getfa(e[i].y);
if (u!=v) link(u,v,col(e[i].x)==col(e[i].y));
else if (col(e[i].x)==col(e[i].y))
{
for(int i=l;i<=r;i++)puts("NO");
RE(now);return;
}
swap(e[k--],e[i--]);
}
if (l==r) puts("YES");
else
{
int i,j;
for (i=1,j=0;i<=k;i++)
if(e[i].l<=mid)swap(e[i],e[++j]);//?????? ??
work(l,mid,j);
for (i=1,j=0;i<=k;i++)
if(e[i].r>mid)swap(e[i],e[++j]);
work(mid+1,r,j);
}
RE(now);//??
}
map<int,int>M[N];
int main()
{ scanf("%d%d",&n,&T);
for (int i=1;i<=T;i++)
{
int x,y;
scanf("%d%d",&x,&y);
if (x>y) swap(x,y);
if (M[x][y]) e[M[x][y]].r=i-1,M[x][y]=0;
else
{
M[x][y]=++m;
e[m].x=x;e[m].y=y;e[m].l=i-1;e[m].r=T;
}
}
for (int i=1;i<=m;i++)
{
// scanf("%d%d%d%d",&e[i].x,&e[i].y,&e[i].l,&e[i].r);
if (++e[i].l>e[i].r) i--,m--;
}
for (int i=1;i<=n;i++)fa[i]=i;
work(1,T,m);
return 0;
}
Educational Codeforces Round 22 补题 CF 813 A-F的更多相关文章
- Educational Codeforces Round 27 补题
题目链接:http://codeforces.com/contest/845 A. Chess Tourney 水题,排序之后判断第n个元素和n+1个元素是不是想等就可以了. #include < ...
- Educational Codeforces Round 23 补题小结
昨晚听说有教做人场,去补了下玩. 大概我的水平能做个5/6的样子? (不会二进制Trie啊,我真菜) A. 傻逼题.大概可以看成向量加法,判断下就好了. #include<iostream> ...
- cordforce Educational Codeforces Round 47 补题笔记 <未完>
题目链接 http://codeforces.com/contest/1009 A. Game Shopping 直接模拟即可,用了一个队列来存储账单 #include <iostream> ...
- Educational Codeforces Round 12补题 经典题 再次爆零
发生了好多事情 再加上昨晚教育场的爆零 ..真的烦 题目链接 A题经典题 这个题我一开始推公式wa 其实一看到数据范围 就算遍历也OK 存在的问题进制错误 .. 思路不清晰 两个线段有交叉 并不是端点 ...
- Educational Codeforces Round 22 E. Army Creation
Educational Codeforces Round 22 E. Army Creation 题意:求区间[L,R]内数字次数不超过k次的这些数字的数量的和 思路:和求区间内不同数字的数量类似,由 ...
- Educational Codeforces Round 22 E. Army Creation(分块好题)
E. Army Creation time limit per test 2 seconds memory limit per test 256 megabytes input standard in ...
- 【Educational Codeforces Round 22】
又打了一场EDU,感觉这场比23难多了啊…… 艹还是我太弱了. A. 随便贪心一下. #include<bits/stdc++.h> using namespace std; ,ans=- ...
- Educational Codeforces Round 22 E. Army Creation 主席树 或 分块
http://codeforces.com/contest/813/problem/E 题目大意: 给出长度为n的数组和k, 大小是1e5级别. 要求在线询问区间[l, r]权值, 权值定义为对于 ...
- Educational Codeforces Round 22 B. The Golden Age(暴力)
题目链接:http://codeforces.com/contest/813/problem/B 题意:就是有一个数叫做不幸运数,满足题目的 n = x^a + y^b,现在给你一个区间[l,r],让 ...
随机推荐
- [USACO11NOV]牛的障碍Cow Steeplechase(匈牙利算法)
洛谷传送门 题目描述: 给出N平行于坐标轴的线段,要你选出尽量多的线段使得这些线段两两没有交点(顶点也算),横的与横的,竖的与竖的线段之间保证没有交点,输出最多能选出多少条线段. 因为横的与横的,竖的 ...
- BZOJ4553 - [TJOI2016]序列
Portal Description 给出一个\(n(n\leq10^5)\)个数的数列\(\{a_n\}\)和\(m(m\leq10^5)\)个形如\((x,y)\)的变化,表示\(a_x\)可以变 ...
- POJ 3615 Cow Hurdles
http://poj.org/problem?id=3615 floyd 最短路径的变形 dist[i][j]变化为 : i j之间的最大边 那么输入的时候可以直接把dist[i][j] 当作i j ...
- android开发里跳过的坑——android studio打包的APK签名无效
近期把一个项目从eclipse上移植到了android studio, 在打包发布APK的时候,应用上传到应用市场时提示取不到签名.但是,我确实使用了 做过签名了. 然后换了一种打包方式 build ...
- 洛谷P2814 家谱(gen)
题目背景 现代的人对于本家族血统越来越感兴趣. 题目描述 给出充足的父子关系,请你编写程序找到某个人的最早的祖先. 输入输出格式 输入格式: 输入由多行组成,首先是一系列有关父子关系的描述,其中每一组 ...
- 远程连接mongodb时,27017端口连接不上的解决办法
一.背景描述: 我在linux RED7上安装了mongodb,并没有修改mongodb的配置文件.然后通过另外一台电脑用pymongo连接mongodb时,报错:timeout. ping IP ...
- python学习之 - re模块
re模块功能:实现字符串匹配. 元字符 描述\ 将下一个字符标记符.或一个向后引用.或一个八进制转义符.例如,“\\n”匹配\n.“\n”匹配换行符.序列“\\”匹配“\”而“\(”则匹配“(”.即相 ...
- poj -1185 炮兵阵地 (经典状压dp)
http://poj.org/problem?id=1185 参考博客:http://poj.org/problem?id=1185 大神博客已经讲的很清楚了,注意存状态的时候是从1开始的,所以初始化 ...
- 【转】Linux下添加新硬盘,分区及挂载
原文:http://blog.chinaunix.net/uid-25829053-id-3067619.html ------------------------------------------ ...
- There is no PasswordEncoder mapped for the id "null"
There is no PasswordEncoder mapped for the id "null" 学习了:https://blog.csdn.net/dream_an/ar ...