Browse · MathNet
PrintChina Western Mathematical Olympiad
China counting and probability
Problem
Assume that is a positive integer, and are nonempty subsets of the set . Prove that there are two disjoint and nonempty subsets and such that
Solution
Proof I We prove by induction for .
When , , the proposition holds.
Suppose it holds for . We consider the case of . Suppose are nonempty subsets of . Let . We will prove the following cases.
Case I There exist such that , then . The proposition is proven.
Case II There exists only one such that . There is no loss of generality in supposing , and that is . Now by the inductive assumption, for , there exist two disjoint subsets and such that We write , . Then and differ at the most by the element . (This can be shown by and the definition of .) In this case, we can make the proposition to hold true by putting into or .
Case III No is empty. Now are nonempty subsets of . By the inductive assumption, we show that, for , there exist disjoint subsets and such that In addition, are also nonempty subsets of . By the inductive assumption, we can show that, for , there exist disjoint subsets and such that Again, we write , , and write , . By using , and the definition of , we see that and differ at the most by the element , and so do and . If or , then the proposition holds. Hence we need only to consider the case when and . There is no loss of generality in supposing , but . Now . After amalgamating the sets occurred repeatedly in and , as well as in and , we get two subsets and of such that where , .
Now, if , then the proposition holds. If there is , we write , , , . And there is no loss of generality in assuming that does not belong to and at the same time, and it does not belong to and at the same time too. Hence there are only two possibilities.
(a) and . If there are two sets in containing , then we take away set from the left side in ④. Now since all elements except in belong to (in view of ③), and there are two sets on the left side in ④ containing . Thus after taking away , the number of elements in does not reduce and ④ is still an equality. In the same way, if there are two sets in containing , then we take away from the right side in ④, and ④ still holds. Of course, if there is only one set in and containing , then after taking away from both sides in ④, it remains to be an equality. (Now, by ② and ③, we can see that the two sides of ④ will not become empty sets.)
(b) and , then . Now after taking away from both sides in ④, the resulting expression is still an equality.
In view of the above operation, we have a method to make the two sets of subscripts and in ④ disjoint. Therefore the proposition holds for .
Proof II Here we need to use a fact from linear algebra that vectors in the -dimensional linear space are linearly dependent.
If element is in set , we write it as 1, otherwise write it as 0. Then corresponds to an -dimensional vector, which is nonzero and contains 0 and 1. We write , where Since are vectors in the -dimensional space, so there exists a group of real numbers, not every one of them to be zero, such that Hence, for , there exist two disjoint and nonempty subsets and such that where , , , (Here, it is essential to put the terms with coefficients greater than zero in ⑤ to one side, and those with coefficients less than zero to another side). We conclude that In fact, if element () belongs to the left side in ⑦, then the -th component of the sum of the vectors from the left side in ⑥ must be greater than zero. Thus it makes the -th component of the sum of the vectors from the right side in ⑥ to be greater than zero. Hence, there is , and its -th component is 1, that is, . Conversely, it is also true, that is, ⑦ holds.
When , , the proposition holds.
Suppose it holds for . We consider the case of . Suppose are nonempty subsets of . Let . We will prove the following cases.
Case I There exist such that , then . The proposition is proven.
Case II There exists only one such that . There is no loss of generality in supposing , and that is . Now by the inductive assumption, for , there exist two disjoint subsets and such that We write , . Then and differ at the most by the element . (This can be shown by and the definition of .) In this case, we can make the proposition to hold true by putting into or .
Case III No is empty. Now are nonempty subsets of . By the inductive assumption, we show that, for , there exist disjoint subsets and such that In addition, are also nonempty subsets of . By the inductive assumption, we can show that, for , there exist disjoint subsets and such that Again, we write , , and write , . By using , and the definition of , we see that and differ at the most by the element , and so do and . If or , then the proposition holds. Hence we need only to consider the case when and . There is no loss of generality in supposing , but . Now . After amalgamating the sets occurred repeatedly in and , as well as in and , we get two subsets and of such that where , .
Now, if , then the proposition holds. If there is , we write , , , . And there is no loss of generality in assuming that does not belong to and at the same time, and it does not belong to and at the same time too. Hence there are only two possibilities.
(a) and . If there are two sets in containing , then we take away set from the left side in ④. Now since all elements except in belong to (in view of ③), and there are two sets on the left side in ④ containing . Thus after taking away , the number of elements in does not reduce and ④ is still an equality. In the same way, if there are two sets in containing , then we take away from the right side in ④, and ④ still holds. Of course, if there is only one set in and containing , then after taking away from both sides in ④, it remains to be an equality. (Now, by ② and ③, we can see that the two sides of ④ will not become empty sets.)
(b) and , then . Now after taking away from both sides in ④, the resulting expression is still an equality.
In view of the above operation, we have a method to make the two sets of subscripts and in ④ disjoint. Therefore the proposition holds for .
Proof II Here we need to use a fact from linear algebra that vectors in the -dimensional linear space are linearly dependent.
If element is in set , we write it as 1, otherwise write it as 0. Then corresponds to an -dimensional vector, which is nonzero and contains 0 and 1. We write , where Since are vectors in the -dimensional space, so there exists a group of real numbers, not every one of them to be zero, such that Hence, for , there exist two disjoint and nonempty subsets and such that where , , , (Here, it is essential to put the terms with coefficients greater than zero in ⑤ to one side, and those with coefficients less than zero to another side). We conclude that In fact, if element () belongs to the left side in ⑦, then the -th component of the sum of the vectors from the left side in ⑥ must be greater than zero. Thus it makes the -th component of the sum of the vectors from the right side in ⑥ to be greater than zero. Hence, there is , and its -th component is 1, that is, . Conversely, it is also true, that is, ⑦ holds.
Techniques
Induction / smoothingVectors