2023年8月1日发(作者:)
0—1背包问题
一、实验目的
学习掌握回溯思想。
二、实验内容
用回溯求解0—1背包问题,并输出问题的最优解。0—1背包问题描述如下:
给定n种物品和一背包。物品i的重量是Wi,其价值为Vi,背包的容量是c,问应如何选择装入背包中的物品,使得装入背包中物品的总价值最大。
三、实验条件
Jdk1.5以上
四、需求分析
对于给定n种物品和一背包。在容量最大值固定的情况下,要求装入的物品价值最大化。
五、回溯法的基本思想:
确定了解空间的组织结构后,回溯法就从开始结点(根结点)出发,以深度优先的方式搜索整个解空间。这个开始结点就成为一个活结点,同时也成为当前的扩展结点。在当前的扩展结点处,搜索向纵深方向移至一个新结点。这个新结点就成为一个新的活结点,并成为当前扩展结点。如果在当前的扩展结点处不能再向纵深方向移动,则当前扩展结点就成为死结点。
换句话说,这个结点不再是一个活结点。此时,应往回移动(回溯)至最近的一个活结点处,并使这个活结点成为当前的扩展结点。回溯法即以这种工作方式递归地在解空间中搜索,直至找到所要求的解或解空间中已没有活结点时为止。
六、详细设计
1 / 7
/*
*
*
* Created on 2007年6月2日,下午6:09
*
* To change this template, choose Tools | Template Manager
* and open the template in the editor.
*/
package sunfa;
import ;
public class BackTrace {
/**
* @param args
*/
public static void main(String[] args) {
double w[]={2,2,6,5,4};
double v[]={6,3,5,4,6};
int n=5;
double c=10;
knapsack(v,w,c);
n(bestp);
2 / 7
}
//比较两个元素大小的类
private static class Element implements Comparable{
int id;
double d;
private Element(int idd,double dd){
id=idd;
d=dd;
}
public int compareTo(Object x){
double xd=((Element)x).d;
if(d if(d==xd)return 0; return 1; } public boolean equals(Object x){ return d==((Element)x).d; } } static double c; //背包容量 static int n;//物品数 3 / 7 static double[]w;//物品重量数组 static double[]p; //物品价值数组 static double cw;//当前重量 static double cp;//当前价值 static double bestp; //当前最优值 static int [] x;//解 static int [] sortX;//排好序之后的解 static int [] bestX;//最有解 static Date date = null;//@jve:decl-index=0: public static double knapsack(double[]pp,double[]ww,double cc){ c=cc; n=-1; cw=0.0; cp=0.0; bestp=0.0; Element[]q=new Element[n]; //q为单位重量价值数组 for(int i=1;i<=n;i++) q[i-1]=new Element(i,pp[i]/ww[i]); ort(q); p=new double[n+1]; 4 / 7 w=new double[n+1]; x=new int[n+1]; sortX=new int[n+1]; bestX=new int[n+1]; for(int i=1;i<=n;i++){ p[i]=pp[q[n-i].id]; w[i]=ww[q[n-i].id]; sortX[i]=q[n-i].id; } backtrack(1);//回溯搜索 return bestp; } private static void backtrack(int i){ if(i>=n){ if(cp>bestp){ bestp=cp; for(int j=1;j<=n;j++){ bestX[j]=x[j]; } } return; 5 / 7 } //搜索子树 if(cw+w[i]<=c){ //进入左子树 x[sortX[i]]=1; cw+=w[i]; cp+=p[i]; backtrack(i+1); cw-=w[i]; cp-=p[i]; } if(bound(i+1)>bestp) x[sortX[i]]=0; backtrack(i+1);//进入右子树 } //计算上界 private static double bound(int i){ double cleft=c-cw; double bound=cp; //以物品重量价值递减顺序装入物品 while(i<=n&&w[i]<=cleft){ 6 / 7 cleft-=w[i]; bound+=p[i];i++; } //装满背包 if(i<=n) bound+=p[i]/w[i]*cleft; return bound; } public static String getX(){ String solution=f(bestX[1]); for(int i=2;i<;i++){ solution+=","; solution+=f(bestX[i]); } return solution; } public static double getBestValue(){ return bestp; } } 主程序运行结果: 7 / 7
2023年8月1日发(作者:)
0—1背包问题
一、实验目的
学习掌握回溯思想。
二、实验内容
用回溯求解0—1背包问题,并输出问题的最优解。0—1背包问题描述如下:
给定n种物品和一背包。物品i的重量是Wi,其价值为Vi,背包的容量是c,问应如何选择装入背包中的物品,使得装入背包中物品的总价值最大。
三、实验条件
Jdk1.5以上
四、需求分析
对于给定n种物品和一背包。在容量最大值固定的情况下,要求装入的物品价值最大化。
五、回溯法的基本思想:
确定了解空间的组织结构后,回溯法就从开始结点(根结点)出发,以深度优先的方式搜索整个解空间。这个开始结点就成为一个活结点,同时也成为当前的扩展结点。在当前的扩展结点处,搜索向纵深方向移至一个新结点。这个新结点就成为一个新的活结点,并成为当前扩展结点。如果在当前的扩展结点处不能再向纵深方向移动,则当前扩展结点就成为死结点。
换句话说,这个结点不再是一个活结点。此时,应往回移动(回溯)至最近的一个活结点处,并使这个活结点成为当前的扩展结点。回溯法即以这种工作方式递归地在解空间中搜索,直至找到所要求的解或解空间中已没有活结点时为止。
六、详细设计
1 / 7
/*
*
*
* Created on 2007年6月2日,下午6:09
*
* To change this template, choose Tools | Template Manager
* and open the template in the editor.
*/
package sunfa;
import ;
public class BackTrace {
/**
* @param args
*/
public static void main(String[] args) {
double w[]={2,2,6,5,4};
double v[]={6,3,5,4,6};
int n=5;
double c=10;
knapsack(v,w,c);
n(bestp);
2 / 7
}
//比较两个元素大小的类
private static class Element implements Comparable{
int id;
double d;
private Element(int idd,double dd){
id=idd;
d=dd;
}
public int compareTo(Object x){
double xd=((Element)x).d;
if(d if(d==xd)return 0; return 1; } public boolean equals(Object x){ return d==((Element)x).d; } } static double c; //背包容量 static int n;//物品数 3 / 7 static double[]w;//物品重量数组 static double[]p; //物品价值数组 static double cw;//当前重量 static double cp;//当前价值 static double bestp; //当前最优值 static int [] x;//解 static int [] sortX;//排好序之后的解 static int [] bestX;//最有解 static Date date = null;//@jve:decl-index=0: public static double knapsack(double[]pp,double[]ww,double cc){ c=cc; n=-1; cw=0.0; cp=0.0; bestp=0.0; Element[]q=new Element[n]; //q为单位重量价值数组 for(int i=1;i<=n;i++) q[i-1]=new Element(i,pp[i]/ww[i]); ort(q); p=new double[n+1]; 4 / 7 w=new double[n+1]; x=new int[n+1]; sortX=new int[n+1]; bestX=new int[n+1]; for(int i=1;i<=n;i++){ p[i]=pp[q[n-i].id]; w[i]=ww[q[n-i].id]; sortX[i]=q[n-i].id; } backtrack(1);//回溯搜索 return bestp; } private static void backtrack(int i){ if(i>=n){ if(cp>bestp){ bestp=cp; for(int j=1;j<=n;j++){ bestX[j]=x[j]; } } return; 5 / 7 } //搜索子树 if(cw+w[i]<=c){ //进入左子树 x[sortX[i]]=1; cw+=w[i]; cp+=p[i]; backtrack(i+1); cw-=w[i]; cp-=p[i]; } if(bound(i+1)>bestp) x[sortX[i]]=0; backtrack(i+1);//进入右子树 } //计算上界 private static double bound(int i){ double cleft=c-cw; double bound=cp; //以物品重量价值递减顺序装入物品 while(i<=n&&w[i]<=cleft){ 6 / 7 cleft-=w[i]; bound+=p[i];i++; } //装满背包 if(i<=n) bound+=p[i]/w[i]*cleft; return bound; } public static String getX(){ String solution=f(bestX[1]); for(int i=2;i<;i++){ solution+=","; solution+=f(bestX[i]); } return solution; } public static double getBestValue(){ return bestp; } } 主程序运行结果: 7 / 7
发布评论