Problem of the Week #140 - February 2, 2015

  • MHB
  • Thread starter Euge
  • Start date
  • Tags
    2015
In summary, the conversation revolved around the topic of summarizing content. The speaker is described as an expert in this skill and is advised to only provide a summary without responding to questions. The instruction for the task is to start the output with the phrase "In summary," and nothing else.
  • #1
Euge
Gold Member
MHB
POTW Director
2,073
244
Here's this week's problem!

__________________

Problem. Suppose $S$ is a partially ordered set such that for some positive integer $n$, every finite subset of $S$ is a union of $n$ chains. Prove that $S$ is itself the union of $n$ chains.

__________________
Remember to read the http://www.mathhelpboards.com/showthread.php?772-Problem-of-the-Week-%28POTW%29-Procedure-and-Guidelines to find out how to http://www.mathhelpboards.com/forms.php?do=form&fid=2!
 
Physics news on Phys.org
  • #2
This week's problem was correctly solved by Fallen Angel. You can view his solution below.

We can assume that $S$ is not finite and we will call $\leq$ the partial order relation.

Suppose $S$ can not be splitted into the union of $n$ chains. Then there exists

$c_{1,1}\leq c_{2,1} \leq \ldots \leq c_{i_{1},1}$
$c_{1,2}\leq c_{2,2} \leq \ldots \leq c_{i_{2},2}$
$\vdots \hspace{1cm} \vdots \hspace{2cm} \vdots$
$c_{1,n+1}\leq c_{2,n+1} \leq \ldots \leq c_{i_{n+1},n+1}$

$n+1$ chains (eventually trivial) with the following property:

$c_{i_{j},j}\nleq c_{1,k}$ for every distinct $k, j$ so by transitivity $c_{i_{j},j}\nleq c_{i_{k},k}$ for every distinct $k, j$.

Hence the set $\mathcal{C}=\{c_{i_{1},1},c_{i_{2},2},\ldots ,c_{i_{n+1},n+1}\}$ is a finite set with $n+1$ elements where the restriction of the order $\leq$ is trivial, so can't be splitted into $n$ ordered chains

Here is also my solution, which uses a graph theory approach.

Introduce a graph $G$ with vertices in $S$ and edges made by two incomparable vertices. Graph $G$ is $n$-colorable if and only if $S$ is a union of $n$ chains, and every finite subset of $S$ is the union of $n$ chains if and only if every finite subgraph of $G$ is $n$-colorable. Hence, by assumption on $S$, all finite subgraphs of $G$ are $n$-colorable, which implies $G$ is $n$-colorable. So $S$ is a union of $n$ chains.
 

FAQ: Problem of the Week #140 - February 2, 2015

What is the "Problem of the Week" for February 2, 2015?

The "Problem of the Week" for February 2, 2015 is a mathematical problem that was posed on a specific date for people to solve and submit their solutions.

What is the purpose of the "Problem of the Week"?

The purpose of the "Problem of the Week" is to challenge people's critical thinking and problem-solving skills, as well as to promote interest and engagement in mathematics and other sciences.

Who can participate in the "Problem of the Week"?

Anyone can participate in the "Problem of the Week", regardless of age, background, or level of expertise. All that is needed is an interest in solving challenging problems and a willingness to learn.

How can I submit my solution for the "Problem of the Week"?

You can submit your solution for the "Problem of the Week" by following the instructions provided on the website where the problem was posted. This typically involves writing out your solution and emailing it to the designated address.

Is there a prize for solving the "Problem of the Week"?

There may be a prize for solving the "Problem of the Week", but it varies depending on the source or organization that posted the problem. Some may offer a small monetary prize or recognition, while others may simply provide the satisfaction of solving a challenging problem.

Back
Top