题意:Given an array of unique integers, each integer is strictly greater than 1.
We make a binary tree using these integers and each number may be used for any number of times.
Each non-leaf node's value should be equal to the product of the values of it's children.
How many binary trees can we make? Return the answer modulo 10 ** 9 + 7.
思路: 这道题和一般动态规划题目的形式不太一样,和建树的过程结合起来,但是不要忘记了动态规划的本质是什么,就是有小问题到大问题的转化,可以将大问题分解为小问题,大问题可以由小问题组成,然后看这道题,很明显根节点和子节点是状态转化的关系;
回到这题,思考c,c.left=a,c.right=b,以c为根节点能够构造的树的个数, 以c的右子节点为根节点能构造的树的个数,是什么关系?
是不是dp[c] = dp[a] * dp[b]呢?因为对于每个a的树形式,和对于每个b的树形式,都是不一样的树;且a、b位交换,也是不一样的树,故正确的状态转移方程应该是:dp[c] = sum{dp[a] * dp[b]}
代码:

class Solution {
public int numFactoredBinaryTrees(int[] A) {
long res = 0;
int N = A.length;
long[] dp = new long[N];
Arrays.fill(dp, 1);
Arrays.sort(A);
Map<Integer, Integer> map = new HashMap<>();
for (int i = 0; i < N; i++) {
map.put(A[i], i);
}
int kmod = 1_000_000_007;
for (int i = 0; i < N; i++) {
for (int j = 0; j < i; j++) {
if (A[i] % A[j] == 0) {
int right = A[i] / A[j];
if (map.containsKey(right)) {
dp[i] = (dp[i] + dp[j] * dp[map.get(right)]) % kmod;
}
}
}
}
for (long num : dp) res += num;
return (int)(res % kmod);
}
}
或者简化一下,直接用map存也是可以的:

class Solution {
public int numFactoredBinaryTrees(int[] A) {
final int kmod = 1_000_000_007;
Map<Integer, Long> map = new HashMap<>(); //dp[]
Arrays.sort(A);
for (int i = 0; i < A.length; i++) {
map.put(A[i], 1L);
for (int j = 0; j < i; j++) {
if (A[i] % A[j] == 0 && map.containsKey((A[i] / A[j]))) {
map.put(A[i], (map.get(A[i]) + map.get(A[j]) * map.get(A[i] / A[j])) % kmod);
}
}
}
long res = 0;
for (long num : map.values()) res += num;
return (int) (res % kmod);
}
}
这题还有个坑是注意答案会很大,应该用long型保存,且最后应该求模