False: every ordered set has a minimum. False: every ordered set has a maximum. False: every ordered set has a maximal element. But: True: every finite ordered set has a maximal element. ************ PROVE IT!!!!!!!!!!!!!!!!!! ***************** "take any x; if x is not maximal, replace x by some y > x; repeat until done" x := any element while (there is some y > x): x:= y end while return x Since the set is finite, you will terminate eventually. Termination function f(x) that (1) is always >= 0 and (2) decreases in every step For example, f(x) := "the number of y that are > x" Observation: If x < y than f(y) <= f(x)-1 because f(x) counts everything that f(y) counts, but it counts more, for example, it counts y. Infinitely larger versus arbitrarily large What is the largest antichain? v Can you find an antichain of size 1000? Look at A_k := { (x,y) | x+y = k } This has size k+1, so they become larger and larger. However, every A_k is finite. { A_1, A_2, A_3, ... } this is a set of sets it is an infinite set every element of it is a finite set of all its elements, which set is largest? this question makes no sense: it contains arbitrarily large sets. R = (N x N, {(a,b) | a divides b} ) True or false: R contains an infinite chain. True: {1,2,4,8,16,...} = {2^i | i >= 0} = the powers of two True or false: R contains an infinite antichain Yes, all the prime numbers. X = { (x,y) in NxN | x >= y} R = (X, { ((a,b),(u,v)) in XxX | a=u, b<=v} ) Does it have an infinite antichain? Yes: { (1,1), (2,2), (3,3), ...} Does it have an infinite chain? Arbitrary, but not infinite (1) arbitrary: of size k? (k,1), (k,2), ..., (k,k) (2) infinite: no. Take an element (u,v) in this chain. The chain has at most u elements: it must be a subset of (u,1), (u,2), ..., (u,u). 3.3: Does NxN have an infinite antichain? It seems not. But how to prove? Latex: $x \preceq y$ Hasse diagram: If R is a partial order, then the Hasse diagram R! is defined as R! := { (x,y) | x < y and there is no z in between, i.e., no z with x < z < y} "The minimum information you have to give me such that I can reconstruct R" Not completely true. What's the Hasse diagram of "<=" on the real numbers? It's EMPTYYYYY! I guess the following is true, but I have not bothered to think about it a lot: THEOREM: If R is a finite ordering, then (R!)^* = R. Furthermore, (R1)^* = the intersection of all relations S with S^* = R. This is math for "It is the smallest relation from which you can recover R"