这是一个最小的ES6版本。扁平化和无函数可以从Lodash中提取。

const flatten = xs =>
    xs.reduce((cum, next) => [...cum, ...next], []);

const without = (xs, x) =>
    xs.filter(y => y !== x);

const permutations = xs =>
    flatten(xs.map(x =>
        xs.length < 2
            ? [xs]
            : permutations(without(xs, x)).map(perm => [x, ...perm])
    ));

结果:

permutations([1,2,3])
// [[1, 2, 3], [1, 3, 2], [2, 1, 3], [2, 3, 1], [3, 1, 2], [3, 2, 1]]

这里有另一种“更递归”的解决方案。

function perms(input) { var data = input.slice(); var permutations = []; var n = data.length; if (n === 0) { return [ [] ]; } else { var first = data.shift(); var words = perms(data); words.forEach(function(word) { for (var i = 0; i < n; ++i) { var tmp = word.slice(); tmp.splice(i, 0, first) permutations.push(tmp); } }); } return permutations; } var str = 'ABC'; var chars = str.split(''); var result = perms(chars).map(function(p) { return p.join(''); }); console.log(result); var output = window.document.getElementById('output'); output.innerHTML = result; <div id="output"></div>

输出:

[ 'ABC', 'BAC', 'BCA', 'ACB', 'CAB', 'CBA' ]

大多数其他答案没有利用新的javascript生成器函数，这是一个完美的解决这类问题。在内存中，一次可能只需要一个排列。此外，我更喜欢生成一系列索引的排列，因为这允许我对每个排列进行索引，并直接跳转到任何特定的排列，以及用于排列任何其他集合。

// ES6 generator version of python itertools [permutations and combinations] const range = function*(l) { for (let i = 0; i < l; i+=1) yield i; } const isEmpty = arr => arr.length === 0; const permutations = function*(a) { const r = arguments[1] || []; if (isEmpty(a)) yield r; for (let i of range(a.length)) { const aa = [...a]; const rr = [...r, ...aa.splice(i, 1)]; yield* permutations(aa, rr); } } console.log('permutations of ABC'); console.log(JSON.stringify([...permutations([...'ABC'])])); const combinations = function*(a, count) { const r = arguments[2] || []; if (count) { count = count - 1; for (let i of range(a.length - count)) { const aa = a.slice(i); const rr = [...r, ...aa.splice(0, 1)]; yield* combinations(aa, count, rr); } } else { yield r; } } console.log('combinations of 2 of ABC'); console.log(JSON.stringify([...combinations([...'ABC'], 2)])); const permutator = function() { const range = function*(args) { let {begin = 0, count} = args; for (let i = begin; count; count--, i+=1) { yield i; } } const factorial = fact => fact ? fact * factorial(fact - 1) : 1; return { perm: function(n, permutationId) { const indexCount = factorial(n); permutationId = ((permutationId%indexCount)+indexCount)%indexCount; let permutation = [0]; for (const choiceCount of range({begin: 2, count: n-1})) { const choice = permutationId % choiceCount; const lastIndex = permutation.length; permutation.push(choice); permutation = permutation.map((cv, i, orig) => (cv < choice || i == lastIndex) ? cv : cv + 1 ); permutationId = Math.floor(permutationId / choiceCount); } return permutation.reverse(); }, perms: function*(n) { for (let i of range({count: factorial(n)})) { yield this.perm(n, i); } } }; }(); console.log('indexing type permutator'); let i = 0; for (let elem of permutator.perms(3)) { console.log(`${i}: ${elem}`); i+=1; } console.log(); console.log(`3: ${permutator.perm(3,3)}`);

有点晚了，但喜欢在这里添加一个稍微优雅的版本。可以是任何数组…

function permutator(inputArr) {
  var results = [];

  function permute(arr, memo) {
    var cur, memo = memo || [];

    for (var i = 0; i < arr.length; i++) {
      cur = arr.splice(i, 1);
      if (arr.length === 0) {
        results.push(memo.concat(cur));
      }
      permute(arr.slice(), memo.concat(cur));
      arr.splice(i, 0, cur[0]);
    }

    return results;
  }

  return permute(inputArr);
}

添加ES6(2015)版本。也不会改变原始输入数组。工作在控制台Chrome…

const permutator = (inputArr) => {
  let result = [];

  const permute = (arr, m = []) => {
    if (arr.length === 0) {
      result.push(m)
    } else {
      for (let i = 0; i < arr.length; i++) {
        let curr = arr.slice();
        let next = curr.splice(i, 1);
        permute(curr.slice(), m.concat(next))
     }
   }
 }

 permute(inputArr)

 return result;
}

所以…

permutator(['c','a','t']);

收益率…

[ [ 'c', 'a', 't' ],
  [ 'c', 't', 'a' ],
  [ 'a', 'c', 't' ],
  [ 'a', 't', 'c' ],
  [ 't', 'c', 'a' ],
  [ 't', 'a', 'c' ] ]

和…

permutator([1,2,3]);

收益率…

[ [ 1, 2, 3 ],
  [ 1, 3, 2 ],
  [ 2, 1, 3 ],
  [ 2, 3, 1 ],
  [ 3, 1, 2 ],
  [ 3, 2, 1 ] ]

这是Heap算法的实现(类似于@le_m算法)，只是它是递归的。

function permute_kingzee(arr,n=arr.length,out=[]) {
    if(n == 1) {
        return out.push(arr.slice());
    } else {
        for(let i=0; i<n; i++) {
            permute_kingzee(arr,n-1, out);
            let j = ( n % 2 == 0 ) ? i : 0;
            let t = arr[n-1];
            arr[n-1] = arr[j];
            arr[j] = t;
        }
        return out;
    }
}

它看起来也快得多:https://jsfiddle.net/3brqzaLe/

对这个问题的大多数回答都使用昂贵的操作，如连续插入和删除数组中的项，或重复复制数组。

相反，这是典型的回溯解决方案:

function permute(arr) {
  var results = [],
      l = arr.length,
      used = Array(l), // Array of bools. Keeps track of used items
      data = Array(l); // Stores items of the current permutation
  (function backtracking(pos) {
    if(pos == l) return results.push(data.slice());
    for(var i=0; i<l; ++i) if(!used[i]) { // Iterate unused items
      used[i] = true;      // Mark item as used
      data[pos] = arr[i];  // Assign item at the current position
      backtracking(pos+1); // Recursive call
      used[i] = false;     // Mark item as not used
    }
  })(0);
  return results;
}

permute([1,2,3,4]); // [  [1,2,3,4], [1,2,4,3], /* ... , */ [4,3,2,1]  ]

由于结果数组将非常大，因此逐个迭代结果而不是同时分配所有数据可能是一个好主意。在ES6中，这可以通过生成器来完成:

function permute(arr) {
  var l = arr.length,
      used = Array(l),
      data = Array(l);
  return function* backtracking(pos) {
    if(pos == l) yield data.slice();
    else for(var i=0; i<l; ++i) if(!used[i]) {
      used[i] = true;
      data[pos] = arr[i];
      yield* backtracking(pos+1);
      used[i] = false;
    }
  }(0);
}

var p = permute([1,2,3,4]);
p.next(); // {value: [1,2,3,4], done: false}
p.next(); // {value: [1,2,4,3], done: false}
// ...
p.next(); // {value: [4,3,2,1], done: false}
p.next(); // {value: undefined, done: true}