Math 386: Combinatorics Announcements
Go to homework | course information | syllabus.
Index
Test I (Tues., Sept. 29)
Covering all of Homework Sets I-IV.
Grading guidelines:
A B C D F
84-100 71-84 56-70 50-55 0-49
Test II (Tues., Oct. 27)
Covering Chapter 5 (omitting 5.6) and Sections 6.1-3 (that's HW Sets V-VII).
Grading guidelines:
A B C D F
87-100 72-86 55-71 50-54 0-49
Test III (Tues., Dec. 8)
Covering Homework Sets IX-XII, i.e., Sections 3.1-2 and Chapter 7 (except 7.3).
-Grading guidelines:
A B C D F
90-100+ 80-89 70-79 65-69 0-64
-
The final exam will cover the whole course, i.e., everything we've done.
Quizzes
- Here is a summary sheet (PDF) (or in PostScript) of basic counting problems and
formulas.
- And here is an informative note (PDF) on magic squares of all orders, odd or even. (Just 5 pages.)
- This is a well illustrated article about the order-3 magic hexagon. Very nicely illustrated.
- Meshalkin's Theorem generalizes Sperner's Theorem. We have an n-element set S. A k-composition of S is an
ordered list (i.e., a sequence) of non-empty subsets, π = (S1 , S2 , ..., Sk) such that S = S1
union S2 union ... union Sk, and the subsets are pairwise disjoint. A k-composition antichain is a set of m different
k-compositions,
π1 = (S1,1 , S1,2 , ..., S1,k) ,
π2 = (S2,1 , S2,2 , ..., S2,k) ,
... ,
πm = (Sm,1 , Sm,2 , ..., Sm,k) ,
with the property that all the "first" sets (those in the first position in each composition – that is, the sets S1,1 ,
S2,1 , ..., Sm,1) are an antichain of S; and similarly all the "second" sets are an antichain; etc. up to the k-th sets.
(There's a trick. "First" sets may repeat, as long as the entire composition doesn't repeat.)
Theorem (Meshalkin). The number of k-compositions in a k-composition antichain (that is, the number m) is no greater than the largest
multinomial coefficient of the form
(n choose n1 n2 ... nk)
where n1 + n2 + ... + nk = n.
There's a precise description of the largest multinomial coefficient, which is a generalization of the unimodality theorem for binomial coefficients
(Theorem 5.3.1). The theorem is:
Theorem (Max Multinomial). For fixed positive integers n and k, the maximum multinomial coefficient of the form
(n choose n1 n2 ... nk)
(where n1 + n2 + ... + nk = n) occurs when every ni = h or h+1, where h = [n/k] (the floor of n/k).
Matt Beck and I found a nice proof of Meshalkin's theorem, similar to the proof of Theorem 5.3.3, which I hope to show you in class.
(See "A shorter, simpler, stronger proof of the Meshalkin-Hochberg-Hirsch bounds on
componentwise antichains", Journal of Combinatorial Theory Series A, volume 100 (2002), pages 196-199.)
Questions:
- Find an example of a 2-composition antichain with general n. How is it related to antichains in the usual sense?
- Find an example of a 3-composition antichain with n = 4. How big can you make m? What is the upper bound given in the theorem? Can you find an
example where m = this upper bound? If not, how close can you get?
- Prove the "Max Multinomial" theorem. (Not too easy, but not too hard. Think how we proved the unimodality of binomial coefficients.)
- Read about crook placements!
- In Exercise 2.16, the theorem is Theorem 2.3.1, not 3.3.1.
- In Exercise 2.28, a "block" is the part of a street between one corner and the next.
- Don't forget that Exercise 6.31 has two parts. One part is to show that a(n,k) has the formula given. The other part is to show that a(n,k) counts the number of ways to choose k children, as specified in the exercise. I'm not sure which part should be done first, but I incline towards doing the second part first and using it to find the formula.
- In Exercise 7.51, the section should be 7.5, not 7.6.
For a project you may try to solve a problem (partially or completely), or write up (in your own words) what you find in the literature (books, journal articles, etc.), or a combination of both. Below is a list of likely project topics.
The grade and credit for a project will depend on how much work you do and how good the project report is.
You may work alone, or make it a group project. If it's a collaboration, everyone gets the project grade.
If you want to do a project, you must get my approval in advance. The deadline for approval is Friday, November 20. The deadline for submitting the final project report is Friday, December 11 (the last day of class) at 4:30.
You should show me a preliminary report before the final version, so I can tell you what improvements it needs (if any).
- P-1. Solve Ch. 1, # 4(b).
- P-2. Suppose a board B (possibly with missing squares) has an even number of squares, and every square has at least two neighboring squares. Is it always possible to cover the board perfectly with dominoes? If not, can you refine the conditions so it is possible? (You may add restrictions, or change them.)
- P-3. Prove or disprove that, in the chess master problem, Application 4 (problem F2), there is a sequence of days
in which s/he plays exactly k games, for every value of k ≤ 77, but not for k > 77. Or at least, prove as much of this as you can, even if
not all.
- P-4. Read up on Meshalkin's theorem and write a complete proof in your own words. (I will give you a couple of short papers to read.)
- P-5. Read up on the topic of "subset sums" (Problems 3.9 and F1). Discuss varying the numbers (number of people, maximum age) to see when two (disjoint) subgroups exist with equal age sums.
- P-6. Solve Problem G3. You might try to find a solution in the literature, and write it up in your own words, or you might be able to solve it yourself.
- P-7. For permutations of Sn2, what is the relationship between the property of having no monotone sequence of length
n+1 (as in Problem G2), and the inversion number or sequence?
Go to homework | course information | syllabus.