| Time Limit: 5000MS | Memory Limit: 131072K | |
| Total Submissions: 48231 | Accepted: 14226 | |
| Case Time Limit: 2000MS | ||
Description
You have N integers, A1, A2, ... , AN. You need to deal with two kinds of operations. One type of operation is to add some given number to each number in a given interval. The other is to ask for the sum of numbers in a given interval.
Input
The first line contains two numbers N and Q. 1 ≤ N,Q ≤ 100000.
The second line contains N numbers, the initial values of A1, A2, ... , AN. -1000000000 ≤ Ai ≤ 1000000000.
Each of the next Q lines represents an operation.
"C a b c" means adding c to each of Aa, Aa+1, ... , Ab. -10000 ≤ c ≤ 10000.
"Q a b" means querying the sum of Aa, Aa+1, ... , Ab.
Output
You need to answer all Q commands in order. One answer in a line.
Sample Input
10 5 1 2 3 4 5 6 7 8 9 10 Q 4 4 Q 1 10 Q 2 4 C 3 6 3 Q 2 4
Sample Output
4 55 9 15
Hint
The sums may exceed the range of 32-bit integers.
题意:
有N(1到100000)个数和M(1到100000)个操作,给出这个N个数的每个数的值Ai(-100000000到1000000000)。操作有两类,第一类Q a b,代表将[a,b]这区间内的数相加输出结果;第二类C a b c,代表将[a,b]区间内的每个数都+c处理。输出每当Q询问时候的结果。
思路:
线段树。延迟标记。将1到N进行线段树处理,用ad域来标记。
AC:
#include<stdio.h>
#define MAX 250000+5
typedef struct
{
__int64 l; //区间起点
__int64 r; //区间终点
__int64 sum; //区间总和
__int64 ad; //标记增加量
}node;
node no[MAX];
__int64 num[MAX];
__int64 s;
void build(__int64 from,__int64 to,__int64 i)
{
__int64 mid=(from+to)/2;
no[i].l=from;
no[i].r=to;
no[i].ad=0; //一开始增加量都为0
if(from==to)
{
no[i].sum=num[from];
return;
}
build(from,mid,2*i);
build(mid+1,to,2*i+1);
no[i].sum=no[i*2].sum+no[i*2+1].sum;
}
void find(__int64 from,__int64 to,__int64 i)
{
__int64 mid=(no[i].l+no[i].r)/2;
if(from==no[i].l&&to==no[i].r)
{
if(!no[i].ad)
{
s+=no[i].sum;
return;
}
else
{
s+=no[i].sum+(no[i].r-no[i].l+1)*no[i].ad;
//这两天一直纠结这个问题
//想着如果在这里往下传标记行不行
//后来发现是忘了维护sum值了,就这点纠结了两天
//RE的原因是如果[1,1]的时候往下传就没有意义了,但是为了确保能下传,数组要开大点
//还有是+=,而不是直接=
//no[i].sum+=(no[i].r-no[i].l+1)*no[i].ad;
//no[2*i].ad+=no[i].ad;
//no[2*i+1].as+=no[i].ad;
//no[i].ad=0;
//其实这里可以不需要作标记下传操作的,仅仅只是好奇一下
return;
}
}
else
{
if(no[i].ad)
{
no[i].sum+=(no[i].r-no[i].l+1)*no[i].ad;
no[2*i].ad+=no[i].ad;
no[2*i+1].ad+=no[i].ad;
no[i].ad=0;
//传给左右儿子之后才能清零
}
if(from>=mid+1) find(from,to,2*i+1);
if(to<=mid) find(from,to,2*i);
if(from<=mid&&to>=mid+1)
{
find(from,mid,2*i);
find(mid+1,to,2*i+1);
}
}
}
void add(__int64 from,__int64 to,__int64 i,__int64 w)
{
__int64 mid=(no[i].l+no[i].r)/2;
if(no[i].l==from&&no[i].r==to)
{
no[i].ad+=w;
//当找到这个区间之后,增量标记域增加
//而不是直接总和增加
return;
}
else
{
if(no[i].ad!=0)
{
no[i].sum+=(no[i].r-no[i].l+1)*no[i].ad;
//当下一次查询的时候,标记域的值ad使总和sum增加
no[2*i].ad+=no[i].ad;
no[2*i+1].ad+=no[i].ad;
//增加完后则往下传
no[i].ad=0;
//传完之后本身标记的值已经使总和sum增加完毕,故变为0
}
no[i].sum+=(to-from+1)*w;
//这里的相加不是标记域所导致的,是所输入的区间所导致的
if(from>=mid+1) add(from,to,2*i+1,w);
if(to<=mid) add(from,to,2*i,w);
if(from<=mid&&to>=mid+1)
{
add(from,mid,2*i,w);
add(mid+1,to,2*i+1,w);
}
//继续查找线段树,找到相同的区间则标记
}
}
int main()
{
__int64 n,m;
scanf("%I64d%I64d",&n,&m);
for(__int64 i=1;i<=n;i++)
scanf("%I64d",&num[i]);
build(1,n,1);
while(m--)
{
char c;
scanf(" %c",&c);
if(c=='Q')
{
__int64 from,to;
scanf("%I64d%I64d",&from,&to);
s=0;
find(from,to,1);
printf("%I64d\n",s);
}
if(c=='C')
{
__int64 from,to;
__int64 w;
scanf("%I64d%I64d%I64d",&from,&to,&w);
add(from,to,1,w);
}
}
return 0;
}
总结:
1.维护ad的同时也要维护sum值,记住;
2.注意数据范围;
本文介绍了一种解决区间更新和查询问题的有效方法——线段树,并通过一个具体题目详细展示了如何利用延迟标记优化线段树操作。
392

被折叠的 条评论
为什么被折叠?



