UMTYMP Advanced Topics – Introduction to Combinatorics                                    April 20th, 2005

 

 

Bona Chapter 8 #26. Find an explicit formula for the numbers an if  if , and a0 = 0.

 

First of all, there is a typo in the statement of the problem.  should be , otherwise there is no way to compute any of the values!

 

Now, since an + 1 depends on nan , as well as on n!, we need to use an exponential generating function. Otherwise, the coefficients will grow too fast for the function to converge anywhere except at x = 0 and we will not be able to find a closed-form formula for the generating function. So let .

 

Next we multiply both sides of the recursion by  and sum for all . (We choose  so that the n + 1 will cancel with the factors of n + 1 appearing in the recursion.) This gives

 

Then, since  we have

 

So .


Bona Chapter 8 #28. Find the exponential generating function D(x) for derangements, defined in Example 7.3. Look for several different ways to obtain D(x).

 

Method #1 Using the recursion Dn + 1 = (n + 1)Dn + (– 1)n + 1, .

 

Multiply both sides of the recursion by  and sum for all . This gives

 

 

Method #2 Using the explicit formula .

 

Multiply both sides of the formula by  and sum for all . This gives

 

 

 

Method #3 Using the combinatorial interpretation of the product of exponential generating functions.

 

The exponential generating function for the number of permutations is . Every permutation consists of a certain set of fixed points, and a derangement on the remaining points. By the Product Rule (for exponential generating functions), the exponential generating function for permutations is the exponential generating function for derangements, D(x), times the exponential generating function for fixed points. There is only one way to “permute” fixed points, so the function for the fixed points is just .

 

This means we have , or .

 

 

Method #4 Using the recursion Dn + 2 = (n + 1)Dn + 1 + (n + 1)Dn, .

 

Multiply both sides of the recursion by  and sum for all . This gives

 

 

Now notice that

 

Similarly,

 

So we have

 

Setting x = 0 gives lnD0 = 0 + ln(1) + C, i.e., 0 = C (recall that D0 = 1).

 

Hence , and so .