我一直在学习递归思考，第一个打动我的自然解决方案如下。一个更简单的问题是找到一个短一个字母的字符串的排列。我将假设，并相信我的每一根纤维，我的函数可以正确地找到一个字符串的排列，比我目前正在尝试的字符串短一个字母。

Given a string say 'abc', break it into a subproblem of finding permutations of a string one character less which is 'bc'. Once we have permutations of 'bc' we need to know how to combine it with 'a' to get the permutations for 'abc'. This is the core of recursion. Use the solution of a subproblem to solve the current problem. By observation, we can see that inserting 'a' in all the positions of each of the permutations of 'bc' which are 'bc' and 'cb' will give us all the permutations of 'abc'. We have to insert 'a' between adjacent letters and at the front and end of each permutation. For example

我们有bc

“a”+“bc”=“abc”

“b”+“a”+“c”=“bac”

“b”+“a”=“b”

对于'cb'我们有

a + b = acb

“c”+“a”+“b”=“cab”

“cb”+“a”=“cb”

下面的代码片段将说明这一点。下面是该代码片段的工作链接。

def main():
    result = []
    for permutation in ['bc', 'cb']:
        for i in range(len(permutation) + 1):
            result.append(permutation[:i] + 'a' + permutation[i:])
    return result


if __name__ == '__main__':
    print(main())

完整的递归解将是。下面是完整代码的工作链接。

def permutations(s):
    if len(s) == 1 or len(s) == 0:
        return s
    _permutations = []
    for permutation in permutations(s[1:]):
        for i in range(len(permutation) + 1):
            _permutations.append(permutation[:i] + s[0] + permutation[i:])
    return _permutations


def main(s):
    print(permutations(s))


if __name__ == '__main__':
    main('abc')

//循环'整个字符数组，并保持'i'作为你的排列的基础，并像你交换[ab, ba]一样继续寻找组合

public class Permutation {
    //Act as a queue
    private List<Character> list;
    //To remove the duplicates
    private Set<String> set = new HashSet<String>();

    public Permutation(String s) {
        list = new LinkedList<Character>();
        int len = s.length();
        for(int i = 0; i < len; i++) {
            list.add(s.charAt(i));
        }
    }

    public List<String> getStack(Character c, List<Character> list) {
        LinkedList<String> stack = new LinkedList<String>();
        stack.add(""+c);
        for(Character ch: list) {
            stack.add(""+ch);
        }

        return stack;
    }

    public String printCombination(String s1, String s2) {
        //S1 will be a single character
        StringBuilder sb = new StringBuilder();
        String[] strArr = s2.split(",");
        for(String s: strArr) {
            sb.append(s).append(s1);
            sb.append(",");
        }       
        for(String s: strArr) {
            sb.append(s1).append(s);
            sb.append(",");
        }

        return sb.toString();
    }

    public void printPerumtation() {
        int cnt = list.size();

        for(int i = 0; i < cnt; i++) {
            Character c = list.get(0);
            list.remove(0);
            List<String> stack = getStack(c, list);

            while(stack.size() > 1) {
                //Remove the top two elements
                String s2 = stack.remove(stack.size() - 1);
                String s1 = stack.remove(stack.size() - 1);
                String comS = printCombination(s1, s2);
                stack.add(comS);
            }

            String[] perms = (stack.remove(0)).split(",");
            for(String perm: perms) {
                set.add(perm);
            }

            list.add(c);
        }

        for(String s: set) {
            System.out.println(s);
        }
    }
}

这是一个C解:

#include <stdio.h>
#include <string.h>
#include <math.h>
#include <stdlib.h>


char* addLetter(char* string, char *c) {
    char* result = malloc(sizeof(string) + 2);
    strcpy(result, string);
    strncat(result, c, 1);
    return result;
}

char* removeLetter(char* string, char *c) {
    char* result = malloc(sizeof(string));
    int j = 0;
    for (int i = 0; i < strlen(string); i++) {
        if (string[i] != *c) {
            result[j++] = string[i];
        }
    }
    result[j] = '\0';

    return result;
}

void makeAnagram(char *anagram, char *letters) {

    if (*letters == '\0') {
        printf("%s\n", anagram);
        return;
    }

    char *c = letters;
    while (*c != '\0') {
        makeAnagram(addLetter(anagram, c),
                    removeLetter(letters, c));
        c++;
    }

}

int main() {

    makeAnagram("", "computer");

    return 0;
}

这里有一个优雅的，非递归的O(n!)解:

public static StringBuilder[] permutations(String s) {
        if (s.length() == 0)
            return null;
        int length = fact(s.length());
        StringBuilder[] sb = new StringBuilder[length];
        for (int i = 0; i < length; i++) {
            sb[i] = new StringBuilder();
        }
        for (int i = 0; i < s.length(); i++) {
            char ch = s.charAt(i);
            int times = length / (i + 1);
            for (int j = 0; j < times; j++) {
                for (int k = 0; k < length / times; k++) {
                    sb[j * length / times + k].insert(k, ch);
                }
            }
        }
        return sb;
    }

让我试着用Kotlin来解决这个问题:

fun <T> List<T>.permutations(): List<List<T>> {
    //escape case
    if (this.isEmpty()) return emptyList()

    if (this.size == 1) return listOf(this)

    if (this.size == 2) return listOf(listOf(this.first(), this.last()), listOf(this.last(), this.first()))

    //recursive case
    return this.flatMap { lastItem ->
        this.minus(lastItem).permutations().map { it.plus(lastItem) }
    }
}

核心概念:将长链表分解成小链表+递归

长答案与示例列表[1,2,3,4]:

即使是一个4种组合的列表，在脑海中列出所有可能的排列已经有点令人困惑了，我们需要做的就是避免这种情况。我们很容易理解如何对大小为0、1和2的列表进行排列，因此我们所需要做的就是将它们分解为这些大小中的任何一个，并将它们正确地组合起来。想象一台头奖机器:这个算法将从右向左旋转，然后写下

当列表大小为0或1时，返回空/列表为1 当列表大小为2时处理(例如[3,4])，并生成2个排列([3,4]& [4,3]) 对于每一项，将其标记为最后一项中的最后一项，并找到列表中其余项目的所有排列。(例如，把[4]放在桌子上，把[1,2,3]重新排列) 现在对它的子元素进行所有的排列，把它自己放回列表的末尾(例如:[1,2,3][，4]，[1,3,2][，4]，[2,3,1][，4]，…)

这是另一个更简单的方法来做一个字符串的排列。

public class Solution4 {
public static void main(String[] args) {
    String  a = "Protijayi";
  per(a, 0);

}

static void per(String a  , int start ) {
      //bse case;
    if(a.length() == start) {System.out.println(a);}
    char[] ca = a.toCharArray();
    //swap 
    for (int i = start; i < ca.length; i++) {
        char t = ca[i];
        ca[i] = ca[start];
        ca[start] = t;
        per(new String(ca),start+1);
    }

}//per

}