如何编写大量的for循环

时间:2011-06-05 03:07:45

标签: java python algorithm math loops

我是编程新手,所以如果我没有正确地提出这个问题,我很抱歉。

我有以下代码:

int sum = 100;
int a1 = 20;
int a2 = 5;
int a3 = 10;
for (int i = 0; i * a1 <= sum; i++) {
    for (int j = 0; i * a1 + j * a2 <= sum; j++) {
        for (int k = 0; i * a1 + j * a2 + k * a3 <= sum; k++) {
            if (i * a1 + j * a2 + k * a3 == sum) {
                System.out.println(i + "," + j + "," + k);
            }
        }
    }   
}

它的基本功能是告诉我a1a2a3的不同组合等于上述总和(在本例中为100)。这工作正常,但我现在正试图将它应用于更大的数据集,我不知道如何在没有手动编程for循环或高级知道我将拥有多少变量的情况下(可能在10到6000之间) )。我基本上有一个sql查询从数组加载该数据。

Java或python(我正在学习两者)是否有办法自动创建嵌套的forif循环?

非常感谢。

5 个答案:

答案 0 :(得分:13)

递归。

这听起来像是你要解决的问题:

  

您当前的示例:20x 1 + 5x 2 + 10x 3 = 100

     

所以一般来说你做的是:A 1 x 1 + A 2 x 2 + .. 。+ A n x n = SUM

     

所以你传入一个常量数组{A 1 ,A 2 ,...,A n }   你要解决{x 1 ,x 2 ,...,x n }

    public void findVariables(int[] constants, int sum, 
                              int[] variables, int n, int result) {
        if (n == constants.length) { //your end condition for the recursion
            if (result == sum) {
                printArrayAsList(variables);
            }
        } else if (result <= sum){ //keep going
            for (int i = 0; result + constants[n]*i <= sum; i++) {
                variables[n] = i;
                findVariables(constants, sum, variables, n+1, result+constants[n]*i);
            }
        }
    }

并打电话给你使用的例子:

    findVariables(new int[] {20, 5, 20}, 100, new int[] {0,0,0}, 0, 0)

答案 1 :(得分:7)

虽然它可能无法扩展,但这是一个非常简单的暴力python解决方案,不需要递归:

import itertools
target_sum = 100
a = 20
b = 5
c = 10
a_range = range(0, target_sum + 1, a)
b_range = range(0, target_sum + 1, b)
c_range = range(0, target_sum + 1, c)
for i, j, k in itertools.product(a_range, b_range, c_range):
    if i + j + k == 100:
        print i, ',', j, ',', k

此外,还有一些方法可以计算任意列表列表的笛卡尔积,而无需递归。 (lol =列表清单)

def product_gen(*lol):
    indices = [0] * len(lol)
    index_limits = [len(l) - 1 for l in lol]
    while indices < index_limits:
        yield [l[i] for l, i in zip(lol, indices)]
        for n, index in enumerate(indices):
            index += 1
            if index > index_limits[n]:
                indices[n] = 0
            else:
                indices[n] = index
                break
    yield [l[i] for l, i in zip(lol, indices)]

如果您只是在学习python,那么您可能不熟悉yield语句或zip函数;在这种情况下,下面的代码将更清晰。

def product(*lol):
    indices = [0] * len(lol)
    index_limits = [len(l) - 1 for l in lol]
    index_accumulator = []
    while indices < index_limits:
        index_accumulator.append([lol[i][indices[i]] for i in range(len(lol))])
        for n, index in enumerate(indices):
            index += 1
            if index > index_limits[n]:
                indices[n] = 0
            else:
                indices[n] = index
                break
    index_accumulator.append([lol[i][indices[i]] for i in range(len(lol))])
    return index_accumulator

通过跳过i + j + k大于sum的值,您在代码中做了一件很聪明的事情。这些都没有。但是有可能修改后两个来做到这一点,但失去了一般性。

答案 2 :(得分:3)

使用Java,一些通用的简单实现至少需要两个类:

某些委托传递给递归算法,因此您可以在执行所在的任何位置接收更新。类似的东西:

public interface IDelegate {
   public void found(List<CombinationFinder.FoundElement> nstack);
}

实施,例如:

public class CombinationFinder {
   private CombinationFinder next;
   private int multiplier;

   public CombinationFinder(int multiplier) {
      this(multiplier, null);
   }
   public CombinationFinder(int multiplier, CombinationFinder next) {
      this.multiplier = multiplier;
      this.next = next;
   }

   public void setNext(CombinationFinder next) {
      this.next = next;
   }

   public CombinationFinder getNext() {
      return next;
   }

   public void search(int max, IDelegate d) {
      Stack<FoundElement> stack = new Stack<FoundElement>();
      this.search(0, max, stack, d);
   }

   private void search(int start, int max, Stack<FoundElement> s, IDelegate d) {
      for (int i=0, val; (val = start + (i*multiplier)) <= max; i++) {
         s.push(i);
         if (null != next) {
            next.search(val, max, s, d);
         } else if (val == max) {
            d.found(s);
         } 
         s.pop();
      }
   } 

   static public class FoundElement {
      private int value;
      private int multiplier;
      public FoundElement(int value, int multiplier) {
         this.value = value;
         this.multiplier = multiplier;
      }
      public int getValue() {
         return value;
      }
      public int getMultiplier() {
         return multiplier;
      }
      public String toString() {
         return value+"*"+multiplier;
      }
   }
}

最后,设置并运行(测试):

CombinationFinder a1 = new CombinationFinder(20);
CombinationFinder a2 = new CombinationFinder(5);
CombinationFinder a3 = new CombinationFinder(10);

a1.setNext(a2);
a2.setNext(a3);

a1.search(100, new IDelegate() {
   int count = 1;
   @Override
   public void found(List<CombinationFinder.FoundElement> nstack) {
      System.out.print("#" + (count++) + " Found : ");
      for (int i=0; i<nstack.size(); i++) {
         if (i>0) System.out.print(" + ");
            System.out.print(nstack.get(i));
         }
         System.out.println();
      }
   }
});

将输出36个解决方案。

使用此概念,您可以拥有任意数量的内部循环,甚至可以根据需要自定义每个循环。您甚至可以毫无问题地重用对象(即:a1.setNext(a1);)。

** 更新 **

仅仅因为我喜欢montysolution,我无法抗拒测试它,这是结果,稍微调整一下。

免责声明 所有积分都会转到monty以获取算法

public class PolynomialSolver {

   private SolverResult delegate;
   private int min = 0;
   private int max = Integer.MAX_VALUE;

   public PolynomialSolver(SolverResult delegate) {
      this.delegate = delegate;
   }

   public SolverResult getDelegate() {
      return delegate;
   }

   public int getMax() {
      return max;
   }

   public int getMin() {
      return min;
   }

   public void setRange(int min, int max) {
      this.min = min;
      this.max = max;
   }

   public void solve(int[] constants, int total) {
      solveImpl(constants, new int[constants.length], total, 0, 0);
   }

   private void solveImpl(int[] c, int[] v, int t, int n, int r) {
      if (n == c.length) { //your end condition for the recursion
         if (r == t) {
            delegate.solution(c, v, t);
         }
      } else if (r <= t){ //keep going
         for (int i=min, j; (i<=max) && ((j=r+c[n]*i)<=t); i++) {
            v[n] = i;
            solveImpl(c, v, t, n+1, j);
         }
      }
   }

   static public interface SolverResult {
      public void solution(int[] constants, int[] variables, int total);
   }

   static public void main(String...args) {

      PolynomialSolver solver = new PolynomialSolver(new SolverResult() {
         int count = 1;
         @Override
         public void solution(int[] constants, int[] variables, int total) {
            System.out.print("#"+(count++)+" Found : ");
            for (int i=0, len=constants.length; i<len; i++) {
               if (i>0) System.out.print(" + ");
               System.out.print(constants[i]+"*"+variables[i]);
            }
            System.out.println(" = " + total);
         }
      });

      // test some constants = total
      solver.setRange(-10, 20);
      solver.solve(new int[] {20, 5, 10}, 100); // will output 162 solutions

   }
}

答案 3 :(得分:2)

可能有一种方法可以将变量放入List并使用递归来消除所有循环。然而,这种强力方法的运行时间随着变量的数量呈指数增长。 (即算法可能无法在我们的生命周期内完成数千个变量的数量)。

有一些关于如何更有效地解决丢番图方程的文章。数论不是我的专业领域,但希望这些将有所帮助。

http://www.wikihow.com/Solve-a-Linear-Diophantine-Equation

答案 4 :(得分:2)

基于@ monty的解决方案,但有一些调整。最后一个常数可以通过除法来确定。

public static void findVariables(int[] constants, int sum) {
    findVariables0(constants, sum, new int[constants.length], 0);
}

private static void findVariables0(int[] constants, int remaining, int[] variables, int n) {
    if(n == constants.length - 1) {
        // solution if the remaining is divisible by the last constant.
        if (remaining % constants[n] == 0) {
            variables[n] = remaining/constants[n];
            System.out.println(Arrays.toString(variables));
        }
    } else {
        for (int i = 0, limit = remaining/constants[n]; i <= limit; i++) {
            variables[n] = i;
            findVariables0(constants, remaining - i * constants[n], variables, n+1);
        }
    }
}

public static void main(String... args) {
    findVariables(new int[] { 5,3,2 }, 100);
}

当我将int更改为double时,我不会在99%的情况下使用float,因为您获得的舍入错误不值得您保存的内存。

public static void findVariables(double[] constants, double sum) {
    findVariables0(constants, sum, new double[constants.length], 0);
}

private static void findVariables0(double[] constants, double remaining, double[] variables, int n) {
    if(n == constants.length - 1) {
        // solution if the remaining is divisible by the last constant.
        if (remaining % constants[n] == 0) {
            variables[n] = remaining/constants[n];
            System.out.println(Arrays.toString(variables));
        }
    } else {
        for (int i = 0, limit = (int) (remaining/constants[n]); i <= limit; i++) {
            variables[n] = i;
            findVariables0(constants, remaining - i * constants[n], variables, n+1);
        }
    }
}

public static void main(String... args) {
    findVariables(new double[]{5.5, 3, 2}, 100);
}

编译并运行良好。