问题定义(动态规划)
如果我们把二叉树看成一个图,父子节点之间的连线看成是双向的,我们姑且定义"距离"为两节点之间边的个数。写一个程序求一棵二叉树中相距最远的两个节点之间的距离。
书上的解法
书中对这个问题的分析是很清楚的,我尝试用自己的方式简短覆述。
计算一个二叉树的最大距离有两个情况:
- 情况A: 路径经过左子树的最深节点,通过根节点,再到右子树的最深节点。
- 情况B: 路径不穿过根节点,而是左子树或右子树的最大距离路径,取其大者。
只需要计算这两个情况的路径距离,并取其大者,就是该二叉树的最大距离。
我也想不到更好的分析方法。
但接着,原文的实现就不如上面的清楚 (源码可从这里下载):
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
|
// 数据结构定义
struct NODE
{
NODE* pLeft; // 左子树
NODE* pRight; // 右子树
int nMaxLeft; // 左子树中的最长距离
int nMaxRight; // 右子树中的最长距离
char chValue; // 该节点的值
};
int nMaxLen = 0;
// 寻找树中最长的两段距离
void FindMaxLen(NODE* pRoot)
{
// 遍历到叶子节点,返回
if(pRoot == NULL)
{
return;
}
// 如果左子树为空,那么该节点的左边最长距离为0
if(pRoot -> pLeft == NULL)
{
pRoot -> nMaxLeft = 0;
}
// 如果右子树为空,那么该节点的右边最长距离为0
if(pRoot -> pRight == NULL)
{
pRoot -> nMaxRight = 0;
}
// 如果左子树不为空,递归寻找左子树最长距离
if(pRoot -> pLeft != NULL)
{
FindMaxLen(pRoot -> pLeft);
}
// 如果右子树不为空,递归寻找右子树最长距离
if(pRoot -> pRight != NULL)
{
FindMaxLen(pRoot -> pRight);
}
// 计算左子树最长节点距离
if(pRoot -> pLeft != NULL)
{
int nTempMax = 0;
if(pRoot -> pLeft -> nMaxLeft > pRoot -> pLeft -> nMaxRight)
{
nTempMax = pRoot -> pLeft -> nMaxLeft;
}
else
{
nTempMax = pRoot -> pLeft -> nMaxRight;
}
pRoot -> nMaxLeft = nTempMax + 1;
}
// 计算右子树最长节点距离
if(pRoot -> pRight != NULL)
{
int nTempMax = 0;
if(pRoot -> pRight -> nMaxLeft > pRoot -> pRight -> nMaxRight)
{
nTempMax = pRoot -> pRight -> nMaxLeft;
}
else
{
nTempMax = pRoot -> pRight -> nMaxRight;
}
pRoot -> nMaxRight = nTempMax + 1;
}
// 更新最长距离
if(pRoot -> nMaxLeft + pRoot -> nMaxRight > nMaxLen)
{
nMaxLen = pRoot -> nMaxLeft + pRoot -> nMaxRight;
}
}
|
这段代码有几个缺点:
- 算法加入了侵入式(intrusive)的资料nMaxLeft, nMaxRight
- 使用了全局变量 nMaxLen。每次使用要额外初始化。而且就算是不同的独立资料,也不能在多个线程使用这个函数
- 逻辑比较复杂,也有许多 NULL 相关的条件测试。
我的尝试
我认为这个问题的核心是,情况A 及 B 需要不同的信息: A 需要子树的最大深度,B 需要子树的最大距离。只要函数能在一个节点同时计算及传回这两个信息,代码就可以很简单:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
|
#include <iostream>
using namespace std;
struct NODE
{
NODE *pLeft;
NODE *pRight;
};
struct RESULT
{
int nMaxDistance;
int nMaxDepth;
};
RESULT GetMaximumDistance(NODE* root)
{
if (!root)
{
RESULT empty = { 0, -1 }; // trick: nMaxDepth is -1 and then caller will plus 1 to balance it as zero.
return empty;
}
RESULT lhs = GetMaximumDistance(root->pLeft);
RESULT rhs = GetMaximumDistance(root->pRight);
RESULT result;
result.nMaxDepth = max(lhs.nMaxDepth + 1, rhs.nMaxDepth + 1);
result.nMaxDistance = max(max(lhs.nMaxDistance, rhs.nMaxDistance), lhs.nMaxDepth + rhs.nMaxDepth + 2);
return result;
}
|
计算 result 的代码很清楚;nMaxDepth 就是左子树和右子树的深度加1;nMaxDistance 则取 A 和 B 情况的最大值。
为了减少 NULL 的条件测试,进入函数时,如果节点为 NULL,会传回一个 empty 变量。比较奇怪的是 empty.nMaxDepth = -1,目的是让调用方 +1 后,把当前的不存在的 (NULL) 子树当成最大深度为 0。
除了提高了可读性,这个解法的另一个优点是减少了 O(节点数目) 大小的侵入式资料,而改为使用 O(树的最大深度) 大小的栈空间。这个设计使函数完全没有副作用(side effect)。
总的来说就是动态规划.....