Java非递归版本，简单地添加元素并在可能的值之间重新分配它们。0被忽略，适用于固定的列表(给定的是您可以使用的)或可重复的数字列表。

import java.util.*;

public class TestCombinations {

    public static void main(String[] args) {
        ArrayList<Integer> numbers = new ArrayList<>(Arrays.asList(0, 1, 2, 2, 5, 10, 20));
        LinkedHashSet<Integer> targets = new LinkedHashSet<Integer>() {{
            add(4);
            add(10);
            add(25);
        }};

        System.out.println("## each element can appear as many times as needed");
        for (Integer target: targets) {
            Combinations combinations = new Combinations(numbers, target, true);
            combinations.calculateCombinations();
            for (String solution: combinations.getCombinations()) {
                System.out.println(solution);
            }
        }

        System.out.println("## each element can appear only once");
        for (Integer target: targets) {
            Combinations combinations = new Combinations(numbers, target, false);
            combinations.calculateCombinations();
            for (String solution: combinations.getCombinations()) {
                System.out.println(solution);
            }
        }
    }

    public static class Combinations {
        private boolean allowRepetitions;
        private int[] repetitions;
        private ArrayList<Integer> numbers;
        private Integer target;
        private Integer sum;
        private boolean hasNext;
        private Set<String> combinations;

        /**
         * Constructor.
         *
         * @param numbers Numbers that can be used to calculate the sum.
         * @param target  Target value for sum.
         */
        public Combinations(ArrayList<Integer> numbers, Integer target) {
            this(numbers, target, true);
        }

        /**
         * Constructor.
         *
         * @param numbers Numbers that can be used to calculate the sum.
         * @param target  Target value for sum.
         */
        public Combinations(ArrayList<Integer> numbers, Integer target, boolean allowRepetitions) {
            this.allowRepetitions = allowRepetitions;
            if (this.allowRepetitions) {
                Set<Integer> numbersSet = new HashSet<>(numbers);
                this.numbers = new ArrayList<>(numbersSet);
            } else {
                this.numbers = numbers;
            }
            this.numbers.removeAll(Arrays.asList(0));
            Collections.sort(this.numbers);

            this.target = target;
            this.repetitions = new int[this.numbers.size()];
            this.combinations = new LinkedHashSet<>();

            this.sum = 0;
            if (this.repetitions.length > 0)
                this.hasNext = true;
            else
                this.hasNext = false;
        }

        /**
         * Calculate and return the sum of the current combination.
         *
         * @return The sum.
         */
        private Integer calculateSum() {
            this.sum = 0;
            for (int i = 0; i < repetitions.length; ++i) {
                this.sum += repetitions[i] * numbers.get(i);
            }
            return this.sum;
        }

        /**
         * Redistribute picks when only one of each number is allowed in the sum.
         */
        private void redistribute() {
            for (int i = 1; i < this.repetitions.length; ++i) {
                if (this.repetitions[i - 1] > 1) {
                    this.repetitions[i - 1] = 0;
                    this.repetitions[i] += 1;
                }
            }
            if (this.repetitions[this.repetitions.length - 1] > 1)
                this.repetitions[this.repetitions.length - 1] = 0;
        }

        /**
         * Get the sum of the next combination. When 0 is returned, there's no other combinations to check.
         *
         * @return The sum.
         */
        private Integer next() {
            if (this.hasNext && this.repetitions.length > 0) {
                this.repetitions[0] += 1;
                if (!this.allowRepetitions)
                    this.redistribute();
                this.calculateSum();

                for (int i = 0; i < this.repetitions.length && this.sum != 0; ++i) {
                    if (this.sum > this.target) {
                        this.repetitions[i] = 0;
                        if (i + 1 < this.repetitions.length) {
                            this.repetitions[i + 1] += 1;
                            if (!this.allowRepetitions)
                                this.redistribute();
                        }
                        this.calculateSum();
                    }
                }

                if (this.sum.compareTo(0) == 0)
                    this.hasNext = false;
            }
            return this.sum;
        }

        /**
         * Calculate all combinations whose sum equals target.
         */
        public void calculateCombinations() {
            while (this.hasNext) {
                if (this.next().compareTo(target) == 0)
                    this.combinations.add(this.toString());
            }
        }

        /**
         * Return all combinations whose sum equals target.
         *
         * @return Combinations as a set of strings.
         */
        public Set<String> getCombinations() {
            return this.combinations;
        }

        @Override
        public String toString() {
            StringBuilder stringBuilder = new StringBuilder("" + sum + ": ");
            for (int i = 0; i < repetitions.length; ++i) {
                for (int j = 0; j < repetitions[i]; ++j) {
                    stringBuilder.append(numbers.get(i) + " ");
                }
            }
            return stringBuilder.toString();
        }
    }
}

样例输入:

numbers: 0, 1, 2, 2, 5, 10, 20
targets: 4, 10, 25

样例输出:

## each element can appear as many times as needed
4: 1 1 1 1 
4: 1 1 2 
4: 2 2 
10: 1 1 1 1 1 1 1 1 1 1 
10: 1 1 1 1 1 1 1 1 2 
10: 1 1 1 1 1 1 2 2 
10: 1 1 1 1 2 2 2 
10: 1 1 2 2 2 2 
10: 2 2 2 2 2 
10: 1 1 1 1 1 5 
10: 1 1 1 2 5 
10: 1 2 2 5 
10: 5 5 
10: 10 
25: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
25: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 
25: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 
25: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 
25: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 
25: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 
25: 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 
25: 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 
25: 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 
25: 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 
25: 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 
25: 1 1 1 2 2 2 2 2 2 2 2 2 2 2 
25: 1 2 2 2 2 2 2 2 2 2 2 2 2 
25: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 5 
25: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 5 
25: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 5 
25: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 5 
25: 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 5 
25: 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 5 
25: 1 1 1 1 1 1 1 1 2 2 2 2 2 2 5 
25: 1 1 1 1 1 1 2 2 2 2 2 2 2 5 
25: 1 1 1 1 2 2 2 2 2 2 2 2 5 
25: 1 1 2 2 2 2 2 2 2 2 2 5 
25: 2 2 2 2 2 2 2 2 2 2 5 
25: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 5 5 
25: 1 1 1 1 1 1 1 1 1 1 1 1 1 2 5 5 
25: 1 1 1 1 1 1 1 1 1 1 1 2 2 5 5 
25: 1 1 1 1 1 1 1 1 1 2 2 2 5 5 
25: 1 1 1 1 1 1 1 2 2 2 2 5 5 
25: 1 1 1 1 1 2 2 2 2 2 5 5 
25: 1 1 1 2 2 2 2 2 2 5 5 
25: 1 2 2 2 2 2 2 2 5 5 
25: 1 1 1 1 1 1 1 1 1 1 5 5 5 
25: 1 1 1 1 1 1 1 1 2 5 5 5 
25: 1 1 1 1 1 1 2 2 5 5 5 
25: 1 1 1 1 2 2 2 5 5 5 
25: 1 1 2 2 2 2 5 5 5 
25: 2 2 2 2 2 5 5 5 
25: 1 1 1 1 1 5 5 5 5 
25: 1 1 1 2 5 5 5 5 
25: 1 2 2 5 5 5 5 
25: 5 5 5 5 5 
25: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 10 
25: 1 1 1 1 1 1 1 1 1 1 1 1 1 2 10 
25: 1 1 1 1 1 1 1 1 1 1 1 2 2 10 
25: 1 1 1 1 1 1 1 1 1 2 2 2 10 
25: 1 1 1 1 1 1 1 2 2 2 2 10 
25: 1 1 1 1 1 2 2 2 2 2 10 
25: 1 1 1 2 2 2 2 2 2 10 
25: 1 2 2 2 2 2 2 2 10 
25: 1 1 1 1 1 1 1 1 1 1 5 10 
25: 1 1 1 1 1 1 1 1 2 5 10 
25: 1 1 1 1 1 1 2 2 5 10 
25: 1 1 1 1 2 2 2 5 10 
25: 1 1 2 2 2 2 5 10 
25: 2 2 2 2 2 5 10 
25: 1 1 1 1 1 5 5 10 
25: 1 1 1 2 5 5 10 
25: 1 2 2 5 5 10 
25: 5 5 5 10 
25: 1 1 1 1 1 10 10 
25: 1 1 1 2 10 10 
25: 1 2 2 10 10 
25: 5 10 10 
25: 1 1 1 1 1 20 
25: 1 1 1 2 20 
25: 1 2 2 20 
25: 5 20 
## each element can appear only once
4: 2 2 
10: 1 2 2 5 
10: 10 
25: 1 2 2 20 
25: 5 20

到目前为止，有很多解决方案，但都是生成然后过滤的形式。这意味着他们可能会在递归路径上花费大量时间，而这些递归路径不会导致解决方案。

这里的解决方案是O(size_of_array * (number_of_sum + number_of_solutions))。换句话说，它使用动态规划来避免列举永远不会匹配的可能解决方案。

为了搞笑，我让这个函数同时使用正数和负数，并让它成为一个迭代器。它适用于Python 2.3+。

def subset_sum_iter(array, target):
    sign = 1
    array = sorted(array)
    if target < 0:
        array = reversed(array)
        sign = -1
    # Checkpoint A

    last_index = {0: [-1]}
    for i in range(len(array)):
        for s in list(last_index.keys()):
            new_s = s + array[i]
            if 0 < (new_s - target) * sign:
                pass # Cannot lead to target
            elif new_s in last_index:
                last_index[new_s].append(i)
            else:
                last_index[new_s] = [i]
    # Checkpoint B

    # Now yield up the answers.
    def recur(new_target, max_i):
        for i in last_index[new_target]:
            if i == -1:
                yield [] # Empty sum.
            elif max_i <= i:
                break # Not our solution.
            else:
                for answer in recur(new_target - array[i], i):
                    answer.append(array[i])
                    yield answer

    for answer in recur(target, len(array)):
        yield answer

这里有一个例子，它与数组和目标一起使用，在其他解决方案中使用的过滤方法实际上永远不会结束。

def is_prime(n):
    for i in range(2, n):
        if 0 == n % i:
            return False
        elif n < i * i:
            return True
    if n == 2:
        return True
    else:
        return False


def primes(limit):
    n = 2
    while True:
        if is_prime(n):
            yield(n)
        n = n + 1
        if limit < n:
            break


for answer in subset_sum_iter(primes(1000), 76000):
    print(answer)

这将在2秒内打印所有522个答案。之前的方法如果能在宇宙当前的生命周期内找到答案，那就太幸运了。(整个空间有2^168 = 3.74144419156711e+50个可能的组合。那需要一段时间。)

解释 我被要求解释代码，但解释数据结构通常更能说明问题。我来解释一下数据结构。

让我们考虑subset_sum_iter([2, 2、3、3、5、5、7、7、-11、11),10)。

在检查点A，我们已经意识到我们的目标是正的，所以符号= 1。我们已经对输入进行了排序，使array =[-11， -7， -5， -3， -2, 2,3,5,7,11]。由于我们经常通过索引访问它，下面是从索引到值的映射:

0: -11
1:  -7
2:  -5
3:  -3
4:  -2
5:   2
6:   3
7:   5
8:   7
9:  11

通过检查点B，我们使用动态规划生成last_index数据结构。它包含什么?

last_index = {    
    -28: [4],
    -26: [3, 5],
    -25: [4, 6],
    -24: [5],
    -23: [2, 4, 5, 6, 7],
    -22: [6],
    -21: [3, 4, 5, 6, 7, 8],
    -20: [4, 6, 7],
    -19: [3, 5, 7, 8],
    -18: [1, 4, 5, 6, 7, 8],
    -17: [4, 5, 6, 7, 8, 9],
    -16: [2, 4, 5, 6, 7, 8],
    -15: [3, 5, 6, 7, 8, 9],
    -14: [3, 4, 5, 6, 7, 8, 9],
    -13: [4, 5, 6, 7, 8, 9],
    -12: [2, 4, 5, 6, 7, 8, 9],
    -11: [0, 5, 6, 7, 8, 9],
    -10: [3, 4, 5, 6, 7, 8, 9],
    -9: [4, 5, 6, 7, 8, 9],
    -8: [3, 5, 6, 7, 8, 9],
    -7: [1, 4, 5, 6, 7, 8, 9],
    -6: [5, 6, 7, 8, 9],
    -5: [2, 4, 5, 6, 7, 8, 9],
    -4: [6, 7, 8, 9],
    -3: [3, 5, 6, 7, 8, 9],
    -2: [4, 6, 7, 8, 9],
    -1: [5, 7, 8, 9],
    0: [-1, 5, 6, 7, 8, 9],
    1: [6, 7, 8, 9],
    2: [5, 6, 7, 8, 9],
    3: [6, 7, 8, 9],
    4: [7, 8, 9],
    5: [6, 7, 8, 9],
    6: [7, 8, 9],
    7: [7, 8, 9],
    8: [7, 8, 9],
    9: [8, 9],
    10: [7, 8, 9]
}

(旁注，它不是对称的，因为条件if 0 < (new_s - target) *符号阻止我们记录超过target的任何内容，在我们的例子中是10。)

这是什么意思?以条目10为例:[7,8,9]。这意味着我们可以得到10的最终和，最后选择的数字在索引7、8或9处。也就是说，最后选择的数字可以是5,7或11。

让我们仔细看看如果我们选择索引7会发生什么。这意味着我们以5结束。因此，在得到下标7之前，我们必须得到10-5 = 5。5的条目为5:[6,7,8,9]。所以我们可以选择指数6，也就是3。虽然我们在第7、8和9处得到了5，但在第7号下标之前我们没有得到5。所以倒数第二个选项是指数6处的3。

现在我们要在下标6之前得到5-3 = 2。条目2是:2:[5,6,7,8,9]。同样，我们只关心下标5的答案因为其他的都发生得太晚了。所以倒数第三个选项是指数5处的2。

最后我们要在下标5之前得到2-2 = 0。条目0表示:0:[- 1,5,6,7,8,9]。同样，我们只关心-1。但是-1不是下标实际上我用它来表示我们已经完成了选择。

我们求出了2+3+5 = 10的解。这是我们打印出来的第一个解。

现在我们来看递归子函数。因为它是在main函数内部定义的，所以它可以看到last_index。

首先要注意的是，它调用的是yield，而不是return。这使它成为一个发电机。当你调用它时，你会返回一个特殊类型的迭代器。当你循环遍历那个迭代器时，你会得到一个它能产生的所有东西的列表。但你是在生成它们时得到它们的。如果它是一个很长的列表，你不把它放在内存中。(有点重要，因为我们可以得到一个很长的列表。)

recur(new_target, max_i)将产生的结果是你可以用数组中最大索引为max_i的元素求和为new_target的所有方法。这就是它的答案:“我们必须在索引max_i+1之前到达new_target。”当然，它是递归的。

因此，recur(target, len(array))是所有使用任意索引到达目标的解。这就是我们想要的。

非常有效的算法，使用我几年前用c++写的表格。

如果你设置PRINT 1，它将打印所有的组合(但它不会使用有效的方法)。

它非常高效，在不到10毫秒的时间内计算了超过10^14个组合。

#include <stdio.h>
#include <stdlib.h>
//#include "CTime.h"

#define SUM 300
#define MAXNUMsSIZE 30

#define PRINT 0


long long CountAddToSum(int,int[],int,const int[],int);
void printr(const int[], int);
long long table1[SUM][MAXNUMsSIZE];

int main()
{
    int Nums[]={3,4,5,6,7,9,13,11,12,13,22,35,17,14,18,23,33,54};
    int sum=SUM;
    int size=sizeof(Nums)/sizeof(int);
    int i,j,a[]={0};
    long long N=0;
    //CTime timer1;

    for(i=0;i<SUM;++i) 
        for(j=0;j<MAXNUMsSIZE;++j) 
            table1[i][j]=-1;

    N = CountAddToSum(sum,Nums,size,a,0); //algorithm
    //timer1.Get_Passd();

    //printf("\nN=%lld time=%.1f ms\n", N,timer1.Get_Passd());
    printf("\nN=%lld \n", N);
    getchar();
    return 1;
}

long long CountAddToSum(int s, int arr[],int arrsize, const int r[],int rsize)
{
    static int totalmem=0, maxmem=0;
    int i,*rnew;
    long long result1=0,result2=0;

    if(s<0) return 0;
    if (table1[s][arrsize]>0 && PRINT==0) return table1[s][arrsize];
    if(s==0)
    {
        if(PRINT) printr(r, rsize);
        return 1;
    }
    if(arrsize==0) return 0;

    //else
    rnew=(int*)malloc((rsize+1)*sizeof(int));

    for(i=0;i<rsize;++i) rnew[i]=r[i]; 
    rnew[rsize]=arr[arrsize-1];

    result1 =  CountAddToSum(s,arr,arrsize-1,rnew,rsize);
    result2 =  CountAddToSum(s-arr[arrsize-1],arr,arrsize,rnew,rsize+1);
    table1[s][arrsize]=result1+result2;
    free(rnew);

    return result1+result2;

}

void printr(const int r[], int rsize)
{
    int lastr=r[0],count=0,i;
    for(i=0; i<rsize;++i) 
    {
        if(r[i]==lastr)
            count++;
        else
        {
            printf(" %d*%d ",count,lastr);
            lastr=r[i];
            count=1;
        }
    }
    if(r[i-1]==lastr) printf(" %d*%d ",count,lastr);

    printf("\n");

}

这类似于硬币更换问题

public class CoinCount 
{   
public static void main(String[] args)
{
    int[] coins={1,4,6,2,3,5};
    int count=0;

    for (int i=0;i<coins.length;i++)
    {
        count=count+Count(9,coins,i,0);
    }
    System.out.println(count);
}

public static int Count(int Sum,int[] coins,int index,int curSum)
{
    int count=0;

    if (index>=coins.length)
        return 0;

    int sumNow=curSum+coins[index];
    if (sumNow>Sum)
        return 0;
    if (sumNow==Sum)
        return 1;

    for (int i= index+1;i<coins.length;i++)
        count+=Count(Sum,coins,i,sumNow);

    return count;       
}
}

c++版本的相同算法

#include <iostream>
#include <list>
void subset_sum_recursive(std::list<int> numbers, int target, std::list<int> partial)
{
        int s = 0;
        for (std::list<int>::const_iterator cit = partial.begin(); cit != partial.end(); cit++)
        {
            s += *cit;
        }
        if(s == target)
        {
                std::cout << "sum([";

                for (std::list<int>::const_iterator cit = partial.begin(); cit != partial.end(); cit++)
                {
                    std::cout << *cit << ",";
                }
                std::cout << "])=" << target << std::endl;
        }
        if(s >= target)
            return;
        int n;
        for (std::list<int>::const_iterator ai = numbers.begin(); ai != numbers.end(); ai++)
        {
            n = *ai;
            std::list<int> remaining;
            for(std::list<int>::const_iterator aj = ai; aj != numbers.end(); aj++)
            {
                if(aj == ai)continue;
                remaining.push_back(*aj);
            }
            std::list<int> partial_rec=partial;
            partial_rec.push_back(n);
            subset_sum_recursive(remaining,target,partial_rec);

        }
}

void subset_sum(std::list<int> numbers,int target)
{
    subset_sum_recursive(numbers,target,std::list<int>());
}
int main()
{
    std::list<int> a;
    a.push_back (3); a.push_back (9); a.push_back (8);
    a.push_back (4);
    a.push_back (5);
    a.push_back (7);
    a.push_back (10);
    int n = 15;
    //std::cin >> n;
    subset_sum(a, n);
    return 0;
}

下面是一个Java版本，它非常适合小N和非常大的目标和，当复杂度O(t*N)(动态解)大于指数算法时。我的版本在中间攻击中使用了一个meet，并进行了一些调整，以降低复杂度，从经典的naive O(n*2^n)降低到O(2^(n/2))。

如果你想在32到64个元素之间的集合中使用这种方法，你应该将表示step函数中当前子集的int改为long，尽管随着集合大小的增加，性能显然会急剧下降。如果你想对一个有奇数个元素的集合使用这个，你应该给这个集合加上一个0，使它成为偶数。

import java.util.ArrayList;
import java.util.List;

public class SubsetSumMiddleAttack {
    static final int target = 100000000;
    static final int[] set = new int[]{ ... };

    static List<Subset> evens = new ArrayList<>();
    static List<Subset> odds = new ArrayList<>();

    static int[][] split(int[] superSet) {
        int[][] ret = new int[2][superSet.length / 2]; 

        for (int i = 0; i < superSet.length; i++) ret[i % 2][i / 2] = superSet[i];

        return ret;
    }

    static void step(int[] superSet, List<Subset> accumulator, int subset, int sum, int counter) {
        accumulator.add(new Subset(subset, sum));
        if (counter != superSet.length) {
            step(superSet, accumulator, subset + (1 << counter), sum + superSet[counter], counter + 1);
            step(superSet, accumulator, subset, sum, counter + 1);
        }
    }

    static void printSubset(Subset e, Subset o) {
        String ret = "";
        for (int i = 0; i < 32; i++) {
            if (i % 2 == 0) {
                if ((1 & (e.subset >> (i / 2))) == 1) ret += " + " + set[i];
            }
            else {
                if ((1 & (o.subset >> (i / 2))) == 1) ret += " + " + set[i];
            }
        }
        if (ret.startsWith(" ")) ret = ret.substring(3) + " = " + (e.sum + o.sum);
        System.out.println(ret);
    }

    public static void main(String[] args) {
        int[][] superSets = split(set);

        step(superSets[0], evens, 0,0,0);
        step(superSets[1], odds, 0,0,0);

        for (Subset e : evens) {
            for (Subset o : odds) {
                if (e.sum + o.sum == target) printSubset(e, o);
            }
        }
    }
}

class Subset {
    int subset;
    int sum;

    Subset(int subset, int sum) {
        this.subset = subset;
        this.sum = sum;
    }
}