比方说，我们有一个输入：

Let's say we've got an input:

10 // saying 1st property should be 10(in total)
10 // saying 2d property should be 10 (in total)
5 // saying theres 5 records below
// (1st property) (2nd property) (cost)
2 5 8 
7 3 10 
4 2 9
4 3 5
8 5 15

在此情况下，输出将看起来像这样：

In this case, the output would look like that:

22 // lowest possible cost
1 3 4 // indexes of records, we've been using (indexing starts with 1)

 2  5  8
 4  2  9
 4  3  5
+---------
 10 10 22

如果有没有实现这些属性为10和10，可能的方式计划将输出-1; 我不知道如何解决背包问题，但我有不知道如何解决这个概率。

If there wasn't possible way to achieve those properties to be 10 and 10, program would output -1; I do know how to solve knapsack problem, however I've got no clue how to solve this prob.

推荐答案

您可以使用的背包问题同样的方法，但不是二维矩阵，你将有一个三维表，每个参数（2约束+指数维度）。

You can use the same approach of knapsack problem, but instead of 2D matrix, you will have a 3D table, a dimension for each parameter (2 constraint + index).

的递推公式将是类似的，但当然这两个参数将被完成。

The recursive formula will be similar, but of course will be done for both parameters.

f(item,cost1,cost2) = max {
               f(item-1,cost1,cost2), 
               f(item,cost1-firstCost[i],cost2-secondCost[i]) + profit[i]
                          }

（基本条款是相似的，但是有一个额外的维度。）

(Base clauses will be similar as well, but with an extra dimension.)