1. Let be a - design with blocks and blocks through every point. Let be any block. Show that the number of blocks that meet (excluding ) is at least
Show that equality holds if and only if any block not disjoint from meets it in a constant number of points.
\begin{proof}
Let .
Consider the tuples such that , .
Fix , then
So number of such triples .
For each , define . Consider the tuples with . Then we have
On the other hand, the number of triples . Hence
Now we finish the proof.
\end{proof}
2. A Latin square of order is a quadruple ( ) where , and are sets of cardinality and is a mapping such that for any and , the equation
has a unique solution , and for any , the same equation has a unique solution Here is an example of order 5 :
Define
for .
- (a) Show that a Latin square of order 6 defines a 2-(36, 15, 6) design with the blocks .
- (b) Show that for any other , the sets do not form a 2-design.
\begin{proof}
a) Define , and
so blocks are -subsets.
For any -subset , suppose that , then one of the following holds.
- If and , then
- either , or
- and , or
- and .
- There are blocks containing the .
- If or , then . WLOG , then
- and , there is solution
- and , there is solution
- , , there is solutions
- Hence there are blocks containing the .
- If and and
- and or , there are solutions for
- and or , there are solutions for
- and , or
- and .
- There are blocks containing the .
Therefore, a Latin square of order 6 defines a 2-(36, 15, 6) design with the blocks .
b) By the argument in a), for any two points, they are contained in either blocks or blocks.
So for any other , the sets do not form a 2-design.
\end{proof}
3. Let be a hyperoval in . Any of the points not in has the property that there are exactly secants of through . Take five points of and split them into
This can be done in ways. The two pairs determine two secants that meet a point not in . The line through and meets in a second point . This defines -tuples on the points of , containing five given points.
- (a) Show that this defines a - design.
- (b) Show for some , a convenient choice of , and the canonical hyperoval that this - design is not simple, that is, it is has repeated blocks.
\begin{proof}
a) Note that contains points.
Define blocks as the -tuples defined above, then .
For each -subset , there are -tuples containing these points, so .
Therefore, this defines a - design.
b) Take , then contains points and each block equals .
Hence each choice of deduces the same -tuples and so this design is not simple.
\end{proof}
4. Show that a symmetric - design does not exist.
\begin{proof}
Recall that
Link to originalb=\lambda{v\choose t}/{k\choose t}.
then we have . Hence .
By
Link to original
- odd has a solution .
if such design exists, then we know has a solution in .
Then and we get .
Then and , we get .
Then and .
This procedure can repeat infinitely many times, so it is impossible.
\end{proof}
5. The Petersen graph has spectrum . Show that it is not possible to find three edge-disjoint copies of the Petersen graph in .
\begin{proof}
Assume that there are three edge-disjoint copies of the Petersen graph, then there are three - matrices such that
Since Peterson graph is -regular, each has an eigenvector with eigenvalue . For any , we have . Then in , .
Let and be eigenspaces of and with eigenvalue , then . Since and , there exists , then
and so , which is impossible by spectrum of Peterson graph.
\end{proof}
6. The line graph of a graph has the edges of as vertices, two vertices adjacent when they share a vertex of .
- (a) Show that the Petersen graph is not the line graph of any graph.
- (b) Show that the minimum eigenvalue of is at least . (Hint: Consider , where is the incidence matrix of .)
\begin{proof}
a) Assume that there exists a graph such that is the Peterson graph.
Then
Hence and so . 很不幸,这是有解的: .
If there exists such that , then contains a subgraph , which is impossible. Hence for all . Then is a union of cycles or a paths. Then for all satisfies . Therefore, the Petersen graph is not the line graph of any graph.
b) Notice that , where is the adjacency matrix of .
Since all eigenvalues of are non-negative, the minimum eigenvalue of is at least .
\end{proof}