不相交集的求并算法(按集合大小求并+按高度求并)

【0】README

0.1)本文总结于 数据结构与算法分析, 但源代码均为原创,旨在实现 不相交集ADT的两个操作:合并集合union+查找集合find;
0.2) 不相交集ADT 的 Introduction , 参见 http://blog.csdn.net/PacosonSWJTU/article/details/49716905


【1】灵巧求并算法——按集合大小求并

1.1)大小求并法定义:上面的Union执行是相当任意的, 通过使第二棵树 成为第一棵树的子树而完成合并;对其的改进是借助任意的方法打破现有关系, 使得总让较小的树成为较大树的子树,我们把这种方法叫做 大小求并法
1.2)可以看到, 如果Union 操作都是按照大小求并的话,那么任何节点的深度均不会超过 logN;
1.3)首先注意节点的深度为0, 然后它的深度随着一次 Union 的结果而增加的时候,该节点则被置于至少是 它以前所在树两倍大的一棵树上;因此,它的深度最多可以增加 logN次;

  • 1.3.2) Find 操作 的运行时间为 O(logN), 而连续M次操作则花费 O(MlogN);
  • 1.3.3) 下图指出在16次Union操作后可能得到这种最坏的树;而且如果所有的Union都对相等大小的树进行, 那么这样的树是会得到的;
    这里写图片描述

1.4)为了实现这种方法, 我们需要记住每个树的大小。由于我们实际上只使用一个数组,因此可以让每个根的数组元素包含它 的树的大小的负值;
1.5)已经证明:若使用按大小求并则连续 M次运算需要 O(M)平均时间, 这是因为 当随机的Union执行时, 整个算法一般只有一些很小的集合(通常是一个元素)与 大 集合 合并;

1.6)souce code + printing

  • 1.6.1)download source code :
    https://github.com/pacosonTang/dataStructure-algorithmAnalysis/blob/master/chapter8/p203_unionBySize.c
  • Source Code Statements:
    • S1)显然,本源代码只对元素的size进行了加,没有减;因为我想的话, 元素只能合并一次,也即是元素C起初合并到了集合A,就不能再次合并到集合B, 元素C就一直属于集合A的子集了;
    • S2)当然,这个代码只是 大致上实现了按大小求并的思想,如果元素C还可以再次合并到其他集合的话,这就涉及到集合根元素的size的加减问题了;需要的话,添加之即可;
  • 1.6.2)souce code at a glance
#include <stdio.h>
#include <malloc.h>

#define ElementType int
#define Error(str) printf("\n error: %s \n",str) 

struct UnionSet;
typedef struct UnionSet* UnionSet;

// we adopt the child-sibling expr
struct UnionSet
{
    int parent;
    int size;
    ElementType value;
};

UnionSet makeEmpty(); 
UnionSet* initUnionSet(int size, ElementType* data);
void printSet(UnionSet* set, int size);
void printArray(ElementType data[], int size);
int find(ElementType index, UnionSet* set);

// initialize the union set 
UnionSet* initUnionSet(int size, ElementType* data)
{
    UnionSet* set;  
    int i;

    set = (UnionSet*)malloc(size * sizeof(UnionSet));
    if(!set)
    {
        Error("out of space, from func initUnionSet");        
        return NULL;
    }   

    for(i=0; i<size; i++)
    {
        set[i] = makeEmpty();
        if(!set[i])
            return NULL;
        set[i]->value = data[i];
    }

    return set;
}

// allocate the memory for the single UnionSet and evaluate the parent and size -1
UnionSet makeEmpty()
{
    UnionSet temp;

    temp = (UnionSet)malloc(sizeof(struct UnionSet));
    if(!temp)
    {
        Error("out of space, from func makeEmpty!");        
        return NULL;
    }

    temp->parent = -1;
    temp->size = 1;
    return temp;
}

// merge set1 and set2 by size
void setUnion(UnionSet* set, int index1, int index2)
{
    //judge whether the index1 or index2 equals to -1 ,also -1 represents the root
    if(index1 != -1)
        index1 = find(index1, set);
    if(index2 != -1)
        index2 = find(index2, set);

    if(set[index1]->size > set[index2]->size)
    {
        set[index2]->parent = index1;
        set[index1]->size += set[index2]->size;
    }
    else
    {
        set[index1]->parent = index2;
        set[index2]->size += set[index1]->size;
    }
} 

//find the root of one set whose value equals to given value
int find(ElementType index, UnionSet* set) 
{
    UnionSet temp;  

    while(1)
    {
        temp = set[index];
        if(temp->parent == -1)
            break;
        index = temp->parent;
    }

    return index;       
}   

int main()
{
    int size;
    UnionSet* unionSet;
    ElementType data[] = {110, 245, 895, 658, 321, 852, 147, 458, 469, 159, 347, 28};

    size = 12;
    printf("\n\t====== test for union set by size ======\n");
    //printf("\n\t=== the initial array is as follows ===\n");
    //printArray(data, size); 

    printf("\n\t=== the init union set are as follows ===\n");
    unionSet = initUnionSet(size, data); // initialize the union set over
    printSet(unionSet, size);

    printf("\n\t=== after union(1,5) + union(2,5) + union(3,4) + union(4,5) ===\n");
    setUnion(unionSet, 1, 5);
    setUnion(unionSet, 2, 5);
    setUnion(unionSet, 3, 4);
    setUnion(unionSet, 4, 5);
    printSet(unionSet, size);

    printf("\n\t=== after union(9,8) + union(7,6) + union(3,6) ===\n");
    setUnion(unionSet, 9, 8);
    setUnion(unionSet, 7, 6);
    setUnion(unionSet, 3, 6);   
    printSet(unionSet, size);

    return 0;
}

void printArray(ElementType data[], int size)
{
    int i;

    for(i = 0; i < size; i++)    
        printf("\n\t data[%d] = %d", i, data[i]);                    
    printf("\n\n");
} 

void printSet(UnionSet* set, int size)
{
    int i;
    UnionSet temp;

    for(i = 0; i < size; i++)
    {       
        temp = set[i];
        printf("\n\t parent[%d] = %d", i, temp->parent);                
    }
    printf("\n");
}
  • 1.6.3)printing result
    这里写图片描述

【2】灵巧求并算法——按集合高度求并

2.1)按高度求并定义: 另外一种方法是按照高度求并(按照高度求并是按大小求并的简单修改)(推荐)
2.2)它同样保证所有的树的深入最多是 O(logN)。我们使得 浅的树 成为深 的树的子树,这是一种平缓算法, 因为只有当两颗相等深度的树求并时树的高度才会增加(此时树的高度增加1)。
2.3)source code + printing result

#include <stdio.h>
#include <malloc.h>

#define ElementType int
#define Error(str) printf("\n error: %s \n",str) 

struct UnionSet;
typedef struct UnionSet* UnionSet;

// we adopt the depth-sibling expr
struct UnionSet
{
    int parent;
    int height;
    ElementType value;
};

UnionSet makeEmpty();
UnionSet* initUnionSet(int depth, ElementType* data);
void printSet(UnionSet* set, int depth);
void printArray(ElementType data[], int depth);
int find(ElementType index, UnionSet* set);

// initialize the union set 
UnionSet* initUnionSet(int size, ElementType* data)
{
    UnionSet* set;  
    int i;

    set = (UnionSet*)malloc(size * sizeof(UnionSet));
    if(!set)
    {
        Error("out of space, from func initUnionSet");        
        return NULL;
    }   

    for(i=0; i<size; i++)
    {
        set[i] = makeEmpty();
        if(!set[i])
            return NULL;
        set[i]->value = data[i];
    }

    return set;
}

// allocate the memory for the single UnionSet and evaluate the parent and depth -1
UnionSet makeEmpty()
{
    UnionSet temp;

    temp = (UnionSet)malloc(sizeof(struct UnionSet));
    if(!temp)
    {
        Error("out of space, from func makeEmpty!");        
        return NULL;
    }

    temp->parent = -1;
    temp->height = 0;
    return temp;
}

// merge set1 and set2 by depth
void setUnion(UnionSet* set, int index1, int index2)
{
    //judge whether the index1 or index2 equals to -1 ,also -1 represents the root
    if(index1 != -1)
        index1 = find(index1, set);
    if(index2 != -1)
        index2 = find(index2, set);

    if(set[index1]->height > set[index2]->height)   
        set[index2]->parent = index1;           
    else if(set[index1]->height < set[index2]->height)  
        set[index1]->parent = index2;       
    else
    {
        set[index1]->parent = index2;
        set[index2]->height++; // only if the height of both of subtrees is equal, the height increases one
    }
} 

//find the root of one set whose value equals to given value
int find(ElementType index, UnionSet* set) 
{
    UnionSet temp;  

    while(1)
    {
        temp = set[index];
        if(temp->parent == -1)
            break;
        index = temp->parent;
    }

    return index;       
}   

int main()
{
    int size;
    UnionSet* unionSet;
    ElementType data[] = {110, 245, 895, 658, 321, 852, 147, 458, 469, 159, 347, 28};

    size = 12;
    printf("\n\t====== test for union set by depth ======\n");
    //printf("\n\t=== the initial array is as follows ===\n");
    //printArray(data, depth); 

    printf("\n\t=== the init union set are as follows ===\n");
    unionSet = initUnionSet(size, data); // initialize the union set over
    printSet(unionSet, size);

    printf("\n\t=== after union(0, 1) + union(2, 3) + union(4, 5) + union(6, 7) + union(8, 9) + union(10 ,11) ===\n");
    setUnion(unionSet, 0, 1);
    setUnion(unionSet, 2, 3);
    setUnion(unionSet, 4, 5);
    setUnion(unionSet, 6, 7);
    setUnion(unionSet, 8, 9);   
    setUnion(unionSet, 10, 11); 
    printSet(unionSet, size);

    printf("\n\t=== after union(1, 3) + union(5, 7) + union(9, 11) ===\n");
    setUnion(unionSet, 1, 3);
    setUnion(unionSet, 5, 7);
    setUnion(unionSet, 9, 11);
    printSet(unionSet, size);  

    printf("\n\t=== after union(3, 7) + union(7, 11) ===\n");
    setUnion(unionSet, 3, 7);
    setUnion(unionSet, 7, 11);  
    printSet(unionSet, size); 

    return 0;
}

void printArray(ElementType data[], int size)
{
    int i;

    for(i = 0; i < size; i++)    
        printf("\n\t data[%d] = %d", i, data[i]);                    
    printf("\n\n");
} 

void printSet(UnionSet* set, int size)
{
    int i;
    UnionSet temp;

    for(i = 0; i < size; i++)
    {       
        temp = set[i];
        printf("\n\t parent[%d] = %d", i, temp->parent);                
    }
    printf("\n");
}
  • 2.3.3)printing results:
    这里写图片描述
  • 1
    点赞
  • 5
    收藏
    觉得还不错? 一键收藏
  • 0
    评论

“相关推荐”对你有帮助么?

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值