How to Generate Parentheses using Bruteforce or Depth First Sear

  • 时间：2020-09-23 15:11:59
  • 分类：网络文摘
  • 阅读：80 次

Given n pairs of parentheses, write a function to generate all combinations of well-formed parentheses. For example, given n = 3, a solution set is:

1
2
3
4
5
6
7

[
  "((()))",
  "(()())",
  "(())()",
  "()(())",
  "()()()"
]
[
  "((()))",
  "(()())",
  "(())()",
  "()(())",
  "()()()"
]

We can use the following two algorithms: bruteforce, or backtracking/Depth First Search Algorithm to generate n pairs of valid Parentheses. We can use recursion to implement both algorithms as the follows.

Bruteforce Algorithm to Generate Parentheses

First, we can check if a given string is a valid Parentheses. If we meet a ‘(‘ we increment the balance and ‘)’ we decrement it. At any time, if the balance is negative, it is invalid. Finally a valid Parentheses should have zero balance at the end.

1
2
3
4
5
6
7
8
9
10
11
12
13
14

bool isValidParentheses(const string &s) {
    int balance = 0;
    for (int i = 0; i < s.size(); ++ i) {
        if (s[i] == '(') {
            balance ++;
        } else {
            balance --;
            if (balance < 0) {
                return false;
            }
        }
    }
    return balance == 0;
}
bool isValidParentheses(const string &s) {
    int balance = 0;
    for (int i = 0; i < s.size(); ++ i) {
        if (s[i] == '(') {
            balance ++;
        } else {
            balance --;
            if (balance < 0) {
                return false;
            }
        }
    }
    return balance == 0;
}

Then, we can use the recursion to bruteforce all the possible strings and check one-by-one if it is a valid using isValidParentheses method. As each character has two possibilities, and there are N*2 characters, there will be O(2^(2N)) time complexity, and over-all algorithm complexity will be O(n*2^(2N)) including the isValidParentheses which takes O(N).

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

class Solution {
public:
    vector<string> generateParenthesis(int n) {
        vector<string> r;        
        string cur(n * 2, ' ');
        dfs(r, cur, 0);
        return r;
    }
    
private:
    void dfs(vector<string> &res, string &cur, int pos) {
        if (pos == cur.size()) {
            if (isValidParentheses(cur)) {
                res.push_back(cur);
            }
        } else {
            cur[pos] = '(';
            dfs(res, cur, pos + 1);
            cur[pos] = ')';
            dfs(res, cur, pos + 1);
        }
        
    }
};
class Solution {
public:
    vector<string> generateParenthesis(int n) {
        vector<string> r;        
        string cur(n * 2, ' ');
        dfs(r, cur, 0);
        return r;
    }
    
private:
    void dfs(vector<string> &res, string &cur, int pos) {
        if (pos == cur.size()) {
            if (isValidParentheses(cur)) {
                res.push_back(cur);
            }
        } else {
            cur[pos] = '(';
            dfs(res, cur, pos + 1);
            cur[pos] = ')';
            dfs(res, cur, pos + 1);
        }
        
    }
};

Backtracking/DFS Algorithm to Generate Parentheses

Apparently, the performance of the bruteforce algorithm is not ideal as we don’t need to generate the invalid Parentheses in the first place. We can use two counters to remember the number of opening and closed Parentheses respectively, and only backtracking those valid branches. The invalid branches are cut-off and abandoned. 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22

class Solution {
public:
    vector<string> generateParenthesis(int n) {
        vector<string> r;        
        dfs(r, 0, 0, n, "");
        return r;
    }
    
private:
    void dfs(vector<string> &res, int open, int close, int n, string cur) {
        if (cur.size() == 2 * n) {
            res.push_back(cur);
            return;
        }
        if (open < n) {
            dfs(res, open + 1, close, n, cur + "(");
        }
        if (close < open) { // the close Parentheses should be less than the opening Parentheses
            dfs(res, open, close + 1, n, cur + ")");
        }
    }
};
class Solution {
public:
    vector<string> generateParenthesis(int n) {
        vector<string> r;        
        dfs(r, 0, 0, n, "");
        return r;
    }
    
private:
    void dfs(vector<string> &res, int open, int close, int n, string cur) {
        if (cur.size() == 2 * n) {
            res.push_back(cur);
            return;
        }
        if (open < n) {
            dfs(res, open + 1, close, n, cur + "(");
        }
        if (close < open) { // the close Parentheses should be less than the opening Parentheses
            dfs(res, open, close + 1, n, cur + ")");
        }
    }
};

The total number of different Parentheses is the n-th Catalan number, which is 1/(n+1)C(2n, n) and the complexity is bounded asymptotically to 4^n/(n*sqrt(n)).

The same algorithm implemented in Python, as follows:

1
2
3
4
5
6
7
8
9
10
11
12
13
14

class Solution:
    def generateParenthesis(self, n: int) -> List[str]:
        arr = []
        def dfs(open, close, n, cur):            
            nonlocal arr
            if len(cur) == 2 * n:
                arr.append(cur)
                return            
            if open < n:
                dfs(open + 1, close, n, cur + '(')
            if close < open:
                dfs(open, close + 1, n, cur + ')')            
        dfs(0, 0, n, '')
        return arr
class Solution:
    def generateParenthesis(self, n: int) -> List[str]:
        arr = []
        def dfs(open, close, n, cur):            
            nonlocal arr
            if len(cur) == 2 * n:
                arr.append(cur)
                return            
            if open < n:
                dfs(open + 1, close, n, cur + '(')
            if close < open:
                dfs(open, close + 1, n, cur + ')')            
        dfs(0, 0, n, '')
        return arr

–EOF (The Ultimate Computing & Technology Blog) —

推荐阅读：
北山移文原文及翻译　 五柳先生传原文及翻译　 桃花源记原文及翻译　 进水排水的牛吃草问题　 汽车运化肥的牛吃草问题　 下面的哪个分数大　 扩大鱼池问题　 巧分食盐水问题　 均分田地问题　 切西瓜的数学问题　
评论列表
添加评论