美文网首页
华为2018编程挑战赛分配部分二维背包c++代码实现

华为2018编程挑战赛分配部分二维背包c++代码实现

作者: 十月石榴2013 | 来源:发表于2018-04-22 21:26 被阅读63次

这个月参加了华为编程挑战赛,(初赛赛题:http://codecraft.devcloud.huaweicloud.com/home/detail)。
分配部分的代码主要由我负责,放出来做个备份。
背包法的参考资料:
https://baike.baidu.com/item/%E8%83%8C%E5%8C%85%E9%97%AE%E9%A2%98/2416931?fr=aladdin
https://blog.csdn.net/na_beginning/article/details/62884939

1. distrubute.h

#ifndef Dis_H
#define Dis_H
#include <string>
#include <iostream>
#include "predict.h"
#include <list>
using namespace std;
#define ISCPU true
#define ISMEM false
#define FLAVORNUMMAX 999
#define LOWISTRATE 0.3
#define LOWRATE 0.95



//int f[FLAVORNUMMAX][CPUSUM+1][MEMSUM+1] = { 0 };  // 存放value的数组
//bool g[FLAVORNUMMAX][CPUSUM+1][MEMSUM+1] = { false }; // 存放是否放入背包的数组


class Flavor
{
public:
    string flavor;  
    int CPU;        
    int Memory;     
    Flavor(const string FLAVOR){
        flavor = FLAVOR;
        str2int(type,FLAVOR);
        switch (type)
        {
        case 1:
            CPU = 1;
            Memory = 1;
            break;
        case 2:
            CPU = 1;
            Memory = 2;
            break;
        case 3:
            CPU = 1;
            Memory = 4;
            break;
        case 4:
            CPU = 2;
            Memory = 2;
            break;
        case 5:
            CPU = 2;
            Memory = 4;
            break;
        case 6:
            CPU = 2;
            Memory = 8;
            break;
        case 7:
            CPU = 4;
            Memory = 4;
            break;
        case 8:
            CPU = 4;
            Memory = 8;
            break;
        case 9:
            CPU = 4;
            Memory = 16;
            break;
        case 10:
            CPU = 8;
            Memory = 8;
            break;
        case 11:
            CPU = 8;
            Memory = 16;
            break;
        case 12:
            CPU = 8;
            Memory = 32;
            break;
        case 13:
            CPU = 16;
            Memory = 16;
            break;
        case 14:
            CPU = 16;
            Memory = 32;
            break;
        case 15:
            CPU = 16;
            Memory = 64;
            break;
        default:
            cout << "ERROR: ÊäÈëµÄÐéÄâ»úÐͺÅûÓÐÕÒµ½" << endl;
        }
    }

    int type;       // 0~15 虚拟机型号
};
class BagItem
{
public:
    string flavor;
    int CPUWeight;
    int memoryWeight;
    int value;  // 该虚拟机的CPU个数或内存大小
    int num; // 有多少种该虚拟机
    BagItem(const Flavor FLAVOR, const int NUM, const bool CM1){
        flavor = FLAVOR.flavor;
        CPUWeight = FLAVOR.CPU * NUM;
        memoryWeight = FLAVOR.Memory * NUM;
        value = (CM1 == ISCPU) ? CPUWeight : memoryWeight;
        num = NUM;
        type = FLAVOR.type;
    };
    int type;       // 0~15 虚拟机型号
};

class PhysicalServer
{
public:
    list<BagItem> flavorList;   // ·物理机上存放的虚拟机列表
    int num;    // ÎïÀí»ú±êºÅ
    PhysicalServer(int NUM)
    {
        num = NUM;
    };
};


list<BagItem> PreTreatmentSingle(const Flavor flavor, const int NUM);
list<BagItem> PreTreatment(list<Flavor> flavorList, int * num);
void InitialFlavorList(list<Flavor> &flavorList);
void GetF(list<BagItem> itemList);
void GetG(list<BagItem> &itemList, list<PhysicalServer> &serverList);
list<PhysicalServer> GetPhysicalServerList(list<BagItem> &itemList, list<PhysicalServer> &serverList);
vector<double> Distribute(list<PhysicalServer> &serverList, int *a,bool CM_in,int CPUSUM_in, int MEMSUM_in);
#endif

2. distrubute.cpp

#include "distribute.h"
#include <iostream>
#include <list>
#include <vector>
#include <exception>
using namespace std;
bool CM;
int CPUSUM; // 物理服务器的cpu总数
int MEMSUM; // 物理服务器的内存总量
//#define LOWISTRATE 0.3

list<BagItem> PreTreatmentSingle(const Flavor flavor, const int NUM )
{

    list<BagItem> itemList;

    if (NUM <= 0)
    {
        return itemList;
    }

    if (NUM == 1)
    {
        BagItem item(flavor, NUM, CM);
        itemList.push_back(item);
        return itemList;
    }

    int n = NUM;
    int k = 1;
    while (n > 1)
    {
        BagItem item(flavor, k, CM);
        itemList.push_back(item);
        k = k * 2;
        n = n / 2;
    }
    BagItem item(flavor, NUM - k + 1, CM);
    itemList.push_back(item);
    return itemList;
}

// 预处理
list<BagItem> PreTreatment(list<Flavor> flavorList, int * num)
{

    list<BagItem> bagItemList;
    int j = 0;
    list<Flavor>::iterator it;
    for(it = flavorList.begin(); it != flavorList.end(); it++)  //for each (auto flavor in flavorList)
    {
        Flavor flavor = *it;
        if (num[j] != 0){
            for (int i = 0; i < num[j]; i++)
            {
                BagItem item(flavor, 1, CM);
                bagItemList.push_back(item);
            }
        }
        j++;
    }
    return bagItemList;

}
void InitialFlavorList(list<Flavor> &flavorList)
{
    for (int i = 1; i <= 15; i++)
    {
        Flavor flavor(to_string(i));
        flavorList.push_back(flavor);
    }
}


int f[FLAVORNUMMAX][100][513] = { 0 };  
bool g[FLAVORNUMMAX][100][513] = { false }; 

list<PhysicalServer> serverList;

void GetF(list<BagItem> itemList)
{
    int i = 0;
    list<BagItem>::iterator it = itemList.begin();
    for (it = itemList.begin(); it != itemList.end();it++)
    {
        BagItem bagItem = *it;
        for (int u = 1; u <= CPUSUM; u++)
        {
            for (int v = 1; v <= MEMSUM; v++)
            {
                if ((i > 0) && (u - bagItem.CPUWeight >= 0) && (v - bagItem.memoryWeight >= 0))      // 一般情况
                {
                    if (f[i - 1][u][v] < f[i - 1][u - bagItem.CPUWeight][v - bagItem.memoryWeight] + bagItem.value)
                    {
                        f[i][u][v] = f[i - 1][u - bagItem.CPUWeight][v - bagItem.memoryWeight] + bagItem.value;
                        g[i][u][v] = true;
                    }
                    else if (f[i - 1][u][v] == f[i - 1][u - bagItem.CPUWeight][v - bagItem.memoryWeight] + bagItem.value)
                    {
                        f[i][u][v] = f[i - 1][u][v];
                        --it;
                        g[i][u][v] = bagItem.value>(*it).value ? true : false;
                        ++it;
                    }
                    else
                    {
                        f[i][u][v] = f[i - 1][u][v];
                        g[i][u][v] = false;
                    }
                }
                else if (i == 0)    // 第一行
                {
                    if ((u - bagItem.CPUWeight >= 0) && (v - bagItem.memoryWeight >= 0)) {
                        f[i][u][v] = bagItem.value;
                        g[i][u][v] = true;
                    }
                    else{
                        f[i][u][v] = 0;
                        g[i][u][v] = false;
                    }
                }
                else        // 重量减了之后超限
                {
                    f[i][u][v] = f[i - 1][u][v];
                    g[i][u][v] = false;
                }
            }
        }
        i++;
    }

}

void GetG(list<BagItem> &itemList, list<PhysicalServer> &serverList)
{
    int u = CPUSUM ;
    int v = MEMSUM ;
    int i = itemList.size() - 1;
    PhysicalServer physicalServer(serverList.size());
    list<BagItem>::iterator it = itemList.end();  // 迭代器初始化为数组最末尾
    for (it--; it != itemList.begin(); it--)        // 逆序访问物品列表
    {
        if (g[i][u][v])
        {
            physicalServer.flavorList.push_back(*it);   // 放在该物理机中
            u = u - (*it).CPUWeight;
            v = v - (*it).memoryWeight;
            it = itemList.erase(it);        // 删除元素, 指向被删除元素之后的元素
        }
        i--;
    }   // 直到it指向itemList.begin,此时i应等于0
    if (g[i][u][v])     // 处理itemList.begin这个元素
    {
        physicalServer.flavorList.push_back(*it);   // 放在该物理机中
        itemList.erase(it);     // 删除元素, 指向被删除元素之后的元素
    }
    serverList.push_back(physicalServer);
}

list<PhysicalServer> GetPhysicalServerList(list<BagItem> &itemList, list<PhysicalServer> &serverList)
{
    while (itemList.size() > 0)
    {
        GetF(itemList);
        GetG(itemList, serverList);
    }
    return serverList;
}

/////////////////////////////////////////////////////////////////////////////////////////////////////

void AddFlavor(list<BagItem> &flavorList, int *a, bool CM_in, int CPUSUM_in, int MEMSUM_in)
{
    int cpuu = 0;
    int mem = 0;
    list<BagItem>::iterator it2;
    for (it2 = flavorList.begin(); it2 != flavorList.end(); it2++)
    {
        cpuu += (*it2).CPUWeight * (*it2).num;
        mem += (*it2).memoryWeight * (*it2).num;
    }

    int cpuLeft =  CPUSUM_in-cpuu;
    int memLeft =  MEMSUM_in-mem;

    // int order[10]={ 13, 12, 10, 9, 7, 6, 4, 3 ,1, 0 };  // flavor14,flavor13, 11, 10...... 2, 1
    // if(CM_in == ISMEM)
    // {
    //  order[0]=14; // flavor15;
    //  order[1]=11; // flavor12;
    //  order[2]=8; // flavor9;
    //  order[3]=7; // flavor8;
    //  order[4]=5; // flavor6;
    //  order[5]=4; // flavor5;
    //  order[6]=2; // flavor3;
    //  order[7]=3; // flavor4;
    //  order[8]=1; // flavor2;
    //  order[9]=0; // flavor1;
    // }
    // int order[7]={ 12, 9, 6, 3, 0, 0, 0 };  // flavor14,flavor13, 11, 10...... 2, 1
    // if(CM_in == ISMEM)
    // {
    //  order[0]=14; // flavor15;
    //  order[1]=11; // flavor12;
    //  order[2]=8; // flavor9;
    //  order[3]=5; // flavor6;
    //  order[4]=2; // flavor3;
    //  order[5]=1; // flavor2;
    //  order[6]=0; // flavor1;
    // }
    int order[10]={9, 7, 6, 13, 12, 10,  4, 3 ,1, 0 };  // flavor14,flavor13, 11, 10...... 2, 1
    if(CM_in == ISMEM)
    {
        order[0]=8; // flavor9;
        order[1]=7; // flavor8;
        order[2]=5; // flavor6;
        order[3]=14; // flavor15;
        order[4]=11; // flavor12;
        order[5]=4; // flavor5;
        order[6]=2; // flavor3;
        order[7]=3; // flavor4;
        order[8]=1; // flavor2;
        order[9]=0; // flavor1;
    }

    for(int i=0;i<10;i++)
    {
        if(a[order[i]]>=0)  // if flavor i  need be predicted 
        {   
            Flavor f(to_string(order[i]+1));
            BagItem flavor(f, 1, CM_in);

            while((cpuLeft >= flavor.CPUWeight)&&(memLeft >= flavor.memoryWeight))
            {
                for (it2 = flavorList.begin(); it2 != flavorList.end(); it2++)
                {
                    if ((*it2).type == flavor.type)
                    {
                        (*it2).num++ ;
                        break;
                    }
                }
                if(it2 == flavorList.end())
                    flavorList.push_back(flavor);
                a[order[i]]++;
                cpuLeft -=  flavor.CPUWeight;
                memLeft -=  MEMSUM_in-mem;
            }

        }
    }

}


vector<double> mend(list<PhysicalServer> &serverList, int *a,bool CM_in,int CPUSUM_in, int MEMSUM_in)
{
    CM = CM_in;
    CPUSUM = CPUSUM_in;
    MEMSUM = MEMSUM_in;


    list<PhysicalServer>::iterator it;
    //cout<<"hhhhhh"<<endl;
    for(it = serverList.begin(); it != serverList.end(); it++)// each (auto server in serverList)
    {
        //PhysicalServer server = *it;
        //cout << server.num << ":" << endl;
        int cpuu = 0;
        int mem = 0;
        list<BagItem>::iterator it2;
        for (it2 = (*it).flavorList.begin(); it2 != (*it).flavorList.end(); it2++)
        {
            cpuu += (*it2).CPUWeight * (*it2).num;
            mem += (*it2).memoryWeight * (*it2).num;
        }

        double cp = cpuu*1.0 / CPUSUM;
        double me = mem*1.0 / MEMSUM;

        if ((CM==ISCPU && cp < LOWISTRATE)||(CM==ISMEM && me < LOWISTRATE))
        {
            for (it2 = (*it).flavorList.begin(); it2 != (*it).flavorList.end(); it2++)
            {
                a[(*it2).type-1] = a[(*it2).type-1] - (*it2).num;
                cout<<(*it2).type<<"::"<<(*it2).num<<endl;
            }
            it = serverList.erase(it);
        }
        else if ((CM==ISCPU && cp < LOWRATE)||(CM==ISMEM && me < LOWRATE))
        {
            AddFlavor( (*it).flavorList, a, CM_in, CPUSUM_in, MEMSUM_in);
        }
    }

    double cpurate = 0;
    double memrate = 0;
    int m = 0;
    for(it = serverList.begin(); it != serverList.end(); it++)// each (auto server in serverList)
    {
        //PhysicalServer server = *it;
        (*it).num = m;
        m++;
        //cout << server.num << ":" << endl;
        int cpuu = 0;
        int mem = 0;
        list<BagItem>::iterator it2;
        for (it2 = (*it).flavorList.begin(); it2 != (*it).flavorList.end(); it2++)
        {
            cpuu += (*it2).CPUWeight * (*it2).num;
            mem += (*it2).memoryWeight * (*it2).num;
        }
        double cp = cpuu*1.0 / CPUSUM;
        double me = mem*1.0 / MEMSUM;
        cpurate = cpurate + cp;
        memrate = memrate + me;
    }
    cout << cpurate / m  << "\t" << memrate/m <<  endl;
    vector<double> score2;
    score2.push_back(cpurate / m);
    score2.push_back(memrate / m );
    for(int i=0;i<15;i++)
        cout<<a[i]<<" ";
    cout<<endl;
// if((CM==ISCPU && cpurate / m< LOWRATE )||(CM==ISMEM && memrate / m < LOWRATE))
//  for(int i=0;i<15;i++)
//      a[i]=10000;
    return score2;
}
/////////////////////////////////////////////////////////////////////////////////////////////////////

vector<double> Distribute(list<PhysicalServer> &serverList, int *a,bool CM_in,int CPUSUM_in, int MEMSUM_in)
{
    for(int i=0; i<FLAVORNUMMAX; i++)
    {
        for(int k=0; k<100; k++)
        {
            for(int j=0; j<513; j++)
            {
                f[i][k][j]=0;
                g[i][k][j]=false;
            }
        }
    }
    serverList.clear();
    CM = CM_in;
    CPUSUM = CPUSUM_in;
    MEMSUM = MEMSUM_in;

    list<Flavor> flavorList;
    InitialFlavorList(flavorList);
    //int a[15] = { 6, 1, 6, 3, 9, 9, 0, 59, 8, 2, 11, 0, 1, 16, 0 };
    list<BagItem> bagItemList = PreTreatment(flavorList, a);
    /*for each (auto var in  bagItemList)
    {
        cout << var.flavor << "\t" << var.num << endl;
    }*/
//  list<PhysicalServer> serverList;
    serverList = GetPhysicalServerList(bagItemList, serverList);
    double cpurate = 0;
    double memrate = 0;
    int m = 0;
    list<PhysicalServer>::iterator it;
    //cout<<"hhhhhh"<<endl;
    for(it = serverList.begin(); it != serverList.end(); it++)// each (auto server in serverList)
    {
        //PhysicalServer server = *it;
        m++;
        //cout << server.num << ":" << endl;
        int n = 0;
        int cpuu = 0;
        int mem = 0;
        list<BagItem>::iterator it2;
        for (it2 = (*it).flavorList.begin(); it2 != (*it).flavorList.end(); it2++)
        {
            list<BagItem>::iterator it20 = it2;
            it20++;
            while ((it20 != (*it).flavorList.end()) && ((*it20).flavor == (*it2).flavor))//(flavor.flavor == (*it2).flavor)
            {
                (*it2).num++;
                it20 = (*it).flavorList.erase(it20);
            }
            //cout << (*it2).flavor << "\t" << (*it2).num << endl;
            n += (*it2).num;
            cpuu += (*it2).CPUWeight * (*it2).num;
            mem += (*it2).memoryWeight * (*it2).num;
        }

        double cp = cpuu*1.0 / CPUSUM;
        double me = mem*1.0 / MEMSUM;
        cpurate = cpurate + cp;
        memrate = memrate + me;
        //cout << n << "\t" << cp << "\t" << me << endl;
        //cout << endl;
    }

    cout << cpurate / m  << "\t" << memrate/m <<  endl;
    //char c;
    //c = getchar();

    
    vector<double> score2 = mend(serverList, a, CM_in, CPUSUM_in, MEMSUM_in);
    //  vector<double> score2;
    // score2.push_back(cpurate / m);
    // score2.push_back(memrate / m );
    return score2;
}

相关文章

  • 华为2018编程挑战赛分配部分二维背包c++代码实现

    这个月参加了华为编程挑战赛,(初赛赛题:http://codecraft.devcloud.huaweicloud...

  • Linux下基于C++的TCP连接demo代码分享(C++,Li

    #C++实现TCP连接 @(C++代码)[网络编程, tcp, C++, C++实现] server.cpp: #...

  • C++的泛型编程

    代码膨胀 C++ 的泛型编程是基于模板实现的,而 C++ 的模板采用的是代码膨胀技术。例如std::list容器,...

  • Android 添加JNI 开发能力

    JNI 与 NDK 区别JNI:JNI是一套编程接口,用来实现Java代码与本地的C/C++代码进行交互;NDK:...

  • 1.1、动态规划

    用一纬数组记录,如果不行就用二维,大部分是二维。1、背包问题01背包:有n件物体,容量为v的背包,使物品价值最大。...

  • OC对象的本质

    ~ Objective-C代码,底层实现?Objective-C底层实现是C\C++代码,C\C++代码转换成汇编...

  • 极客班第一周学习笔记

    初识C++ C++是在C之上基于对象,面向对象的编程语言。c++相比c在编程上更加模块化,具象化。 C++代码规范...

  • C++

    C与C++的关系 1.C++可以与C代码进行混编(C++里能写C代码,C里不能写C++代码)2.C++面向对象编程...

  • 剑指Offer.C++.code1-5

    剑指offer 代码实现 C++剑指Offer 1. 二维数组中的查找 2. 替换空格 3. 从尾到头打印链表 4...

  • Swift 添加两种遮罩的方式

    没有遮罩前的效果 代码实现 遮罩方式一 代码实现 遮罩方式二 代码实现 实战 -- 二维码扫描区域遮罩 代码实现 ...

网友评论

      本文标题:华为2018编程挑战赛分配部分二维背包c++代码实现

      本文链接:https://www.haomeiwen.com/subject/coxhlftx.html