一.问题引入
给定n个重量为(w1,w2,…,wn)、价值为{V1,V2,…,Vn)的物品和一个容l量是为C的背包,0/1背包问题(0/1 knapsack problcm)是求这些物品中的一个最有价值的子集,并且要能够装到背包中。
二.算法实例
用蛮力法解决0/1背包问题,需要考虑给定n个物品集合的所有子集,找出所有总重量不超过背包容量的子集,计算每个可能子集的总价值,然后找到价值最大的子集。例如,给定4个物品的重量为(7,3,4,5),价值为{42,12,40,25),和一个容量为10的背包,下表给出了蛮力法求解0/1背包问题的过程。
三.蛮力法过程
序号 | 子集 | 总重量(weight) | 总价值(value) |
1 | 空集 | 0 | 0 |
2 | {1} | 7 | 42 |
3 | {2} | 3 | 12 |
4 | {3} | 4 | 40 |
5 | {4} | 5 | 25 |
6 | {1,2} | 10 | 54 |
7 | {1,3} | 11 | 82 |
8 | {1,4} | 12 | 67 |
9 | {2,3} | 7 | 52 |
10 | {2,4} | 8 | 37 |
11 | {3,4} | 9 | 65 |
12 | {1,2,3} | 14 | 94 |
13 | {1,2,4} | 15 | 79 |
14 | {1,3,4} | 16 | 107 |
15 | {2,3,4} | 12 | 94 |
16 | {1,2,3,4} | 19 | 119 |
四.算法实现(C++)
#define _CRT_SECURE_NO_WARNINGS
#include <stdio.h>
#include <math.h>
int main()
{
int num, maxv = 0;
int n, c, * w, * v, tempw, tempv;
int i, j, k;
printf("input the number and the volume:");
scanf("%d%d", &n, &c);
w = new int[n];
v = new int[n];
printf("input the weights:");
for (i = 0; i < n; i++)
scanf("%d", &w[i]);
printf("input the values:");
for (i = 0; i < n; i++)
scanf("%d", &v[i]);
printf("Selected items:\n");
for (num = 0; num < pow(2, n); num++) //每一个num对应一个解
{
k = num;
tempw = tempv = 0;
printf("Solution %d: ", num);
for (i = 0; i < n; i++) //n位二进制
{
if (k % 2 == 1)
{ //如果相应的位等于1,则代表物体放进去,如果是0,就不用放了
tempw += w[i];
tempv += v[i];
printf("%d ", i);
}
k = k / 2; //二进制转换的规则
}
printf("Total weight: %d,Total value: %d\n", tempw,tempv);
//循环结束后,一个解空间生成,
//判断是否超过了背包的容积,
//如果没有超,判断当前解是否比最优解更好
if (tempw <= c)
{
if (tempv > maxv)
maxv = tempv;
}
}
printf("The maxvalue is %d.\n", maxv);
delete[] w;
delete[] v;
return 0;
}
结果实现:
五.算法实现(Python)
import math
n = int(input("input the number: "))
c = int(input("input the volume: "))
w = []
v = []
print("input the weights:")
for i in range(n):
w.append(int(input()))
print("input the values:")
for i in range(n):
v.append(int(input()))
maxv = 0
print("Selected items:")
for num in range(int(math.pow(2, n))):
k = num
tempw = tempv = 0
print(f"Solution {num}: ", end='')
for i in range(n):
if k % 2 == 1:
tempw += w[i]
tempv += v[i]
print(f"{i} ", end='')
k //= 2
print(f"(Total weight: {tempw},Total value: {tempv})")
if tempw <= c and tempv > maxv:
maxv = tempv
print(f"The maxvalue is {maxv}.")
结果实现:
六.算法实现(java)
package suanfa;
import java.util.Scanner;
public class suanfa {
public static void main(String[] args) {
Scanner scanner = new Scanner(System.in);
System.out.print("input the number: ");
int n = scanner.nextInt();
System.out.print("input the volume: ");
int c = scanner.nextInt();
int[] w = new int[n];
int[] v = new int[n];
System.out.println("input the weights:");
for (int i = 0; i < n; i++) {
w[i] = scanner.nextInt();
}
System.out.println("input the values:");
for (int i = 0; i < n; i++) {
v[i] = scanner.nextInt();
}
int maxv = 0;
System.out.println("Selected items:");
for (int num = 0; num < Math.pow(2, n); num++) {
int k = num;
int tempw = 0;
int tempv = 0;
System.out.print("Solution " + num + ": ");
for (int i = 0; i < n; i++) {
if (k % 2 == 1) {
tempw += w[i];
tempv += v[i];
System.out.print(i + " ");
}
k /= 2;
}
System.out.println("(Total weight: " + tempw + ","+"Total value: "+tempv+")");
if (tempw <= c && tempv > maxv) {
maxv = tempv;
}
}
System.out.println("The maxvalue is " + maxv + ".");
}
}
结果实现:
七.算法分析
对于一个具有n个元素的集合,其子集数量是2^n,所以,不论生成子集的算法效率有多高,蛮力法求解0/1背包问题都会导致一个(2^n)的算法。