691. Stickers to Spell Word

问题：

给定一堆单词卡，每个单词可以有无限张同样的卡片。

要从这些单词卡中选择一些，进行剪切拼接，要构成目标单词，求需要的最少卡片数量。

Example 1:
Input:
["with", "example", "science"], "thehat"
Output:
3
Explanation:
We can use 2 "with" stickers, and 1 "example" sticker.
After cutting and rearrange the letters of those stickers, we can form the target "thehat".
Also, this is the minimum number of stickers necessary to form the target string.

Example 2:
Input:
["notice", "possible"], "basicbasic"
Output:
-1
Explanation:
We can't form the target "basicbasic" from cutting letters from the given stickers.

Note:
stickers has length in the range [1, 50].
stickers consists of lowercase English words (without apostrophes).
target has length in the range [1, 15], and consists of lowercase English letters.
In all test cases, all words were chosen randomly from the 1000 most common US English words, and the target was chosen as a concatenation of two random words.
The time limit may be more challenging than usual. It is expected that a 50 sticker test case can be solved within 35ms on average.

　　

解法：DFS（深度优先搜索）

• 当前状态：
  • 当前目标串target，所需要的最少卡片数。
• 选择：
  • 在所有单词卡中选择，选择后，剩下的字符再进行递归匹配，得到最少卡片数+1，为当前选择的结果。
  • 在这些结果的有效结果（剩下字符递归结果!=-1）中，求res=最小值min
  • 将最小值最为本次递归的结果返回 return res，同时保存在结果map：dp中：dp[target]=res。
• 退出递归条件：
  • dp中保存过，当前目标串target的结果，直接返回dp[target]
  • 当前target串=""，则返回0。（或者提前保存dp[""]=0）

 

♻️ 优化：

Optimization: If the target can be spelled out by a group of stickers, at least one of them has to contain character target[0]. So I explicitly require next sticker containing target[0], which significantly reduced the search space.

// 这个优化剪枝很神奇，可以避免潜在的对target完全没有贡献的sticker，暂时无法贡献的sticker等到其他sticker把部分target拼凑出来以后再贡献

代码参考：

class Solution {
public:
    //dp[s]: for string s, the min number of stickers need.
    unordered_map<string, int> dp;
    int dfs(vector<vector<int>>mp, string target) {
        //return already gotten result.
        if(dp.count(target)) return dp[target];
        //if not already got, then following process:
        int res = INT_MAX;
        //get alphbet map of this target string.
        vector<int>tgmp(26,0);
        for(char a:target) {
            tgmp[a-'a']++;
        }
        //opt: check every sticker
        for(int j=0; j<mp.size(); j++) {
            //every alphbet of this sticker
            //optimization:
            //If the target can be spelled out by a group of stickers, 
            //at least one of them has to contain character target[0]. 
            //So I explicitly require next sticker containing target[0], 
            //which significantly reduced the search space.
            if(mp[j][target[0]-'a']==0)continue;
            //process
            string nexttarget;
            for(int i=0; i<26; i++) {
                if(tgmp[i]>mp[j][i]) {
                    int leftAlphbetCount = tgmp[i]-mp[j][i];
                    nexttarget += string(leftAlphbetCount, 'a'+i);
                }
            }
            // if (nexttarget == target) continue; 
            // 如果没有之前那个优化，一定要判断新target和原target是否一样，避免重复递归
            int tmp = dfs(mp, nexttarget);
            if(tmp!=-1) res = min(res, tmp+1);
        }
        dp[target] = (res==INT_MAX?-1:res);
        return dp[target];
    }
    int minStickers(vector<string>& stickers, string target) {
        //get alphbet map of every word
        int sn = stickers.size();
        vector<vector<int>>mp(sn, vector<int>(26, 0));
        for(int i=0; i<sn; i++) {
            for(char a:stickers[i]) {
                mp[i][a-'a']++;
            }
        }
        dp[""]=0;
        return dfs(mp, target);
    }
};