Bessie and the rest of Farmer John's cows are taking a trip this winter to go skiing. One day Bessie finds herself at the top left corner of an R (1 <= R <= 100) by C (1 <= C <= 100) grid of elevations E (-25 <= E <= 25). In order to join FJ and the other cows at a discow party, she must get down to the bottom right corner as quickly as she can by travelling only north, south, east, and west.

Bessie starts out travelling at a initial speed V (1 <= V <= 1,000,000). She has discovered a remarkable relationship between her speed and her elevation change. When Bessie moves from a location of height A to an adjacent location of eight B, her speed is multiplied by the number 2^(A-B). The time it takes Bessie to travel from a location to an adjacent location is the reciprocal of her speed when she is at the first location.

Find the both smallest amount of time it will take Bessie to join her cow friends.

Input

* Line 1: Three space-separated integers: V, R, and C, which respectively represent Bessie's initial velocity and the number of rows and columns in the grid.

* Lines 2..R+1: C integers representing the elevation E of the corresponding location on the grid.

Output

A single number value, printed to two exactly decimal places: the minimum amount of time that Bessie can take to reach the bottom right corner of the grid.

Sample Input

1 3 3
1 5 3
6 3 5
2 4 3

Sample Output

29.00

Hint

Bessie's best route is: 
Start at 1,1 time 0 speed 1 
East to 1,2 time 1 speed 1/16 
South to 2,2 time 17 speed 1/4 
South to 3,2 time 21 speed 1/8 
East to 3,3 time 29 speed 1/4
 
这个题用vector的话会T
下面是vector是T的代码:
#include<cstdio>
#include<iostream>
#include<cstring>
#include<algorithm>
#include<queue>
#include<stack>
#include<set>
#include<map>
#include<vector>
#include<cmath>
#define Inf 0x3f3f3f3f const int maxn=1e5+;
typedef long long ll;
using namespace std;
ll ksm(ll x,ll y)
{
ll ans=;
while(y)
{
if(y&)
{
ans=ans*x;
}
x*=x;
y>>=;
}
return ans;
}
struct node
{
int to;
double w;
bool friend operator < (node x,node y)
{
return x.w>y.w;
}
};
int Map[]; vector<node>vec[];
double dis[];
int vis[];
int V,R,C;
int dir[][]={{,},{,-},{-,},{,}};
bool check(int x,int y)
{
if(x>=&&x<=R&&y>=&&y<=C)
{
return true;
}
else
{
return false;
}
}
void init()
{
for(int t=;t<=R*C;t++)
{
dis[t]=;
}
} void Dijkstra(int s)
{
node st;
st.to=s;
st.w=;
priority_queue<node>q;
q.push(st);
dis[s]=;
while(!q.empty())
{
node now=q.top();
q.pop();
if(vis[now.to])continue;
vis[now.to]=; int len=vec[now.to].size();
for(int t=;t<len;t++)
{
node tto=vec[now.to][t]; if(vis[tto.to]==&&tto.w+dis[now.to]<dis[tto.to])
{
tto.w=tto.w+dis[now.to];
dis[tto.to]=tto.w;
q.push(tto);
}
}
}
} int main()
{
// std::ios::sync_with_stdio(false);
scanf("%d%d%d",&V,&R,&C);
init();
for(int t=;t<=R;t++)
{
for(int j=;j<=C;j++)
{
scanf("%d",&Map[(t-)*C+j]);
}
}
for(int t=;t<=R;t++)
{
for(int j=;j<=C;j++)
{
for(int k=;k<;k++)
{
int xx=t+dir[k][];
int yy=j+dir[k][];
if(check(xx,yy))
{
node s;
s.to=(xx-)*C+yy;
s.w=1.0/V*ksm(,Map[(t-)*C+j]-Map[]);
vec[(t-)*C+j].push_back(s);
}
}
}
}
Dijkstra();
printf("%.2f\n",dis[R*C]); return ;
} AC的是邻接表的
#include<cstdio>
#include<iostream>
#include<cstring>
#include<algorithm>
#include<queue>
#include<stack>
#include<set>
#include<map>
#include<vector>
#include<cmath> const int maxn=1e4+;
typedef long long ll;
using namespace std; struct edge
{
int u,v;
double w;
int next;
}edge[maxn*]; struct node
{
int pos;
double w;
node(int x,double y)
{
pos=x;
w=y;
}
bool friend operator <(node x,node y)
{
return x.w>y.w;
}
};
int n,m,s,x,y,z,tot = ,V;
bool check(int x,int y)
{
if(x>=&&x<=n&&y>=&&y<=m)
{
return true;
}
else
{
return false;
}
}
int head[];
double d[];
int vis[];
int a[];
int dist[][] = {{-,},{,-},{,},{,}}; void add(int u,int v,double w)
{
edge[++tot].u=u;
edge[tot].v=v;
edge[tot].w=w;
edge[tot].next=head[u];
head[u]=tot;
return ;
} void Dijkstra(int s)
{
priority_queue<node>q;
d[s]=;
q.push(node(s,));
while(!q.empty())
{
node now=q.top();
q.pop();
//cout<<now.pos<<endl;
if(vis[now.pos])continue;
vis[now.pos]=; for(int i=head[now.pos];i!=-;i=edge[i].next)
{
int ne=edge[i].v;
double ww=edge[i].w;
if(d[now.pos]+ww<d[ne])
{
d[ne]=d[now.pos]+ww;
q.push(node(ne,d[ne]));
}
}
}
return ;
}
int main()
{
scanf("%d%d%d",&V,&n,&m);
memset(head,-,sizeof(head));
memset(vis,,sizeof(vis)); for(int t=;t<=n*m;t++)
{
d[t]=;
}
for(int i = ;i <= n; ++i)
for(int j = ;j <= m; ++j)
scanf("%d",&a[(i-)*m+j]);
for(int i = ;i <= n; ++i)
for(int j = ;j <= m; ++j)
for(int k = ;k < ; ++k)
{
int x = i + dist[k][];
int y = j + dist[k][];
if(check(x,y))
{
double v = 1.0 / V * pow(2.0 , a[(i-)*m+j] - a[]);
add((i-)*m+j,(x-)*m+y,v);
}
}
Dijkstra();
printf("%.2f\n",d[n * m]);
return ;
}

POJ - 3037-Skiing(邻接表+Dijkstra)的更多相关文章

  1. POJ 1511 - Invitation Cards 邻接表 Dijkstra堆优化

    昨天的题太水了,堆优化跑的不爽,今天换了一个题,1000000个点,1000000条边= = 试一试邻接表 写的过程中遇到了一些问题,由于习惯于把数据结构封装在 struct 里,结果 int [10 ...

  2. poj 1511(SPFA+邻接表)

    题目链接:http://poj.org/problem?id=1511 思路:题目意思很简单就是要求源点到各点的最短路之和,然后再求各点到源点的最短路之和,其实就是建两个图就ok了,其中一个建反图.1 ...

  3. USACO 2008 January Silver Telephone Lines /// 二分最短路 邻接表dijkstra oj22924

    题目大意: 一共有N (1 ≤ N ≤ 1,000)个电线杆,有P P (1 ≤ P ≤ 10,000)对电线杆是可以连接的, 用几条线连接在一起的电线杆之间都可相互通信,现在想要使得电线杆1和电线杆 ...

  4. POJ - 3255 SPFA+邻接表求次短路径

    题意:给出m条边 , n个顶点,u [ i ]到v [ i ] 的距离w [ i ],求除了最短路的那条最短的边的长度. 思路:之前有做过相似的题,使用迪杰斯特拉算法求单源最短路径,并且记录路径,枚举 ...

  5. POJ 3037 Skiing(Dijkstra)

    Skiing Time Limit: 1000MS   Memory Limit: 65536K Total Submissions: 4668   Accepted: 1242   Special ...

  6. POJ 3037 Skiing(如何使用SPFA求解二维最短路问题)

    题目链接: https://cn.vjudge.net/problem/POJ-3037 Bessie and the rest of Farmer John's cows are taking a ...

  7. POJ - 3037 Skiing SPFA

    Skiing Bessie and the rest of Farmer John's cows are taking a trip this winter to go skiing. One day ...

  8. Poj(2679),SPFA,邻接表(主流写法)

    题目链接:http://poj.org/problem?id=3268 题意: 有编号为1-N的牛,它们之间存在一些单向的路径.给定一头牛的编号,其他牛要去拜访它并且拜访完之后要返回自己原来的位置,求 ...

  9. hdu1839之二分+邻接表+Dijkstra+队列优化

    Delay Constrained Maximum Capacity Path Time Limit: 10000/10000 MS (Java/Others)    Memory Limit: 65 ...

随机推荐

  1. 【av68676164(p41-p42)】内存管理功能

    存储器的功能需求 容量足够大 速度足够快 信息永久保存 多道程序并行 多道程序并行带来的问题 共享:代码和数据共享,节省内存 保护:不允许内存中的程序相互间非法访问 实际存储器体系 三级存储体系 Ca ...

  2. Ant Design Pro V5 从服务器请求菜单(typescript版)

    [前言] 找了很多Admin模板,最后还是看中了AntDesignPro(下文简写antd pro)这个阿里巴巴开源的Admin框架,长这样(还行吧,目前挺主流的): 官网地址:https://pro ...

  3. C#LeetCode刷题之#35-搜索插入位置(Search Insert Position)

    问题 该文章的最新版本已迁移至个人博客[比特飞],单击链接 https://www.byteflying.com/archives/3979 访问. 给定一个排序数组和一个目标值,在数组中找到目标值, ...

  4. K均值聚类和DBSCAN介绍

    K均值(K-means)聚类 问题定义:给定数据$\vec{x}_1,\vec{x}_2,\cdots,\vec{x}_n$,将它们分到不同的$K$个簇(cluster)中.定义$\vec{c}=(c ...

  5. 关于vector的erase删除操作的两种不同方法,在linux与visual studio的实现讨论

    关于vector的erase删除操作的两种不同方法,在linux与visual studio的实现讨论 1.前言: 最近在做某一个题时,用到了vector的删除操作,利用的是erase()函数删除符合 ...

  6. python智能图片识别系统(图片切割、图片识别、区别标识)

    @ 目录 技术介绍 运行效果 关键代码 写在最后 技术介绍 你好! python flask图片识别系统使用到的技术有:图片背景切割.图片格式转换(pdf转png).图片模板匹配.图片区别标识. 运行 ...

  7. 遗传算法框架-基于java jenetics库实现

    本篇并非介绍如何从0开始开发遗传算法框架,反而推荐各位使用已有的GA库jenetics来做遗传算法. GA算法的逻辑还是贴下: 好了,下面介绍的是基于jenetics开发的更贴近业务侧的框架,以及使用 ...

  8. Qt信号与槽使用方法最完整总结

    在图形界面编程中(参考<C++最好的图形库是什么?>),组件之间如何实现通信是核心的技术内容.Qt 使用了信号与槽的机制,非常的高效.简单.易学,方便开发者的使用.本文详细的介绍了Qt 当 ...

  9. 用WEB方式开发WPF桌面程序

    因为疫情影响,公司裁员,结束了一年多的web开发经历,重新开始做桌面,新公司用的是WPF(居然用的是winform style...),当然这跟本文没有关系...上篇博客写的用后台api和前台浏览器控 ...

  10. java多线程:线程间通信——生产者消费者模型

    一.背景 && 定义 多线程环境下,只要有并发问题,就要保证数据的安全性,一般指的是通过 synchronized 来进行同步. 另一个问题是,多个线程之间如何协作呢? 我们看一个仓库 ...