Aizu - ALDS1_5_A Exhaustive Search 【dp】

Write a program which reads a sequence A of n elements and an integer M, and outputs "yes" if you can make M by adding elements in A, otherwise "no". You can use an element only once.

You are given the sequence A and q questions where each question contains Mi.

Input

In the first line n is given. In the second line, n integers are given. In the third line q is given. Then, in the fourth line, q integers (Mi) are given.

Output

For each question Mi, print yes or no.

Constraints

• n ≤ 20
• q ≤ 200
• 1 ≤ elements in A ≤ 2000
• 1 ≤ Mi ≤ 2000

Sample Input 1

5
1 5 7 10 21
8
2 4 17 8 22 21 100 35

Sample Output 1

no
no
yes
yes
yes
yes
no
no

Notes

You can solve this problem by a Burte Force approach. Suppose solve(p, t) is a function which checkes whether you can make t by selecting elements after p-th element (inclusive). Then you can recursively call the following functions:

solve(0, M) 
solve(1, M-{sum created from elements before 1st element}) 
solve(2, M-{sum created from elements before 2nd element}) 
...

The recursive function has two choices: you selected p-th element and not. So, you can check solve(p+1, t-A[p]) and solve(p+1, t) in solve(p, t) to check the all combinations.

For example, the following figure shows that 8 can be made by A[0] + A[2].

 题目大意：给出长度为n的数列，再输入q个数m，判断A中任意几个元素相加能否得到m。如果可以输出yes否则输出no

穷举搜索法：

用函数solve（i，m）代表用第i个数后边的数能得到m时返回true，可将solve（i，m）再分解为solve（i+1，m）和solve（i+1，m-A[i]）。其中solve（i+1，m-A[i]）的意思是使用了第i个数A[i]，由此递归实现

solve（i，m）中当m为0时，代表用现有的数已经凑出来了m；若m>0,但n个数均已用,则说明用现有的数无法凑出m

一个小bug：因为我的程序是从第一个数开始而不是第0个数开始，所以判断不能凑出时应该条件时i>=n+1,这句话说明所有的n个数都已被用

#include<iostream>

using namespace std;

int n;
int A[23];
int solve(int i,int m){
	if(m==0){
		return 1;
	}
	else if(i>=n+1){
		return 0;
	} 
	else{
		int res=solve(i+1,m)||solve(i+1,m-A[i]);
		return res;
	}
}
int main(){
	int q,m;
	cin>>n;
	for(int i=1;i<=n;i++){
		cin>>A[i];
	}
	cin>>q;
	while(q--){
		cin>>m;
		if(solve(1,m)){
			cout<<"yes"<<endl;
		}
		else{
			cout<<"no"<<endl;
		}
	}
	return 0;
} 

动态规划

将算出来的solve（i，m）存入dp[i][m]中，避免重复计算

#include<iostream>
#include<cstring>
using namespace std;

int dp[23][2050];
int n;
int A[23];
int solve(int i,int m){
	if(m==0){
		dp[i][m]=1;
	}
	else if(i>=n+1){
		dp[i][m]=0;
	} 
	else{
		int res=solve(i+1,m)||solve(i+1,m-A[i]);
		dp[i][m]=res;
	}
//    else if(solve(i+1,m)==1) {
//    	dp[i][m]=1;
//	}
//	else if(solve(i+1,m-A[i])==1){
//		dp[i][m]=1;
//	}
//	else{
//		dp[i][m]=0;
//	}
	return dp[i][m];
}
int main(){
	memset(dp,0,sizeof(dp));
	int q,m;
	cin>>n;
	for(int i=1;i<=n;i++){
		cin>>A[i];
	}
	cin>>q;
	while(q--){
		cin>>m;
		if(solve(1,m)==1){
			cout<<"yes"<<endl;
		}
		else{
			cout<<"no"<<endl;
		}
	}
	return 0;
}