This week, we learn about Invariance and Monovariance Principle, which are useful methods to help determine if certain states can be achieved in problems involving sequences, recursions, or iterative processes.
How would you use Invariance and Monovariance to solve the following?
(2008 Putnam) Start with a finite sequence of positive integers. If possible, choose two indices such that does not divide , and replace and by and , respectively. Prove that if this process is repeated, it must eventually stop.
Find many many different ways to approach this question using the ideas of invariance and monovariance.In Worked Example 1, can you classify all possible values which Grayson could be left with on the blackboard? We have already shown that it must be even.
Easy Math Editor
This discussion board is a place to discuss our Daily Challenges and the math and science related to those challenges. Explanations are more than just a solution — they should explain the steps and thinking strategies that you used to obtain the solution. Comments should further the discussion of math and science.
When posting on Brilliant:
*italics*
or_italics_
**bold**
or__bold__
paragraph 1
paragraph 2
[example link](https://brilliant.org)
> This is a quote
\(
...\)
or\[
...\]
to ensure proper formatting.2 \times 3
2^{34}
a_{i-1}
\frac{2}{3}
\sqrt{2}
\sum_{i=1}^3
\sin \theta
\boxed{123}
Comments
Well, I would perhaps use monovariance (for the first problem) in another way, noting first that gcd(aj,ak)×lcm(aj,ak)=ajak. So the whole sequence is bounded above by this product. Then perhaps we can proceed by looking at the last element. Remark that it can only increase since it is replaced by its lcm and some other number we don't really care. So in a way, once it reaches its max value it becomes fixed. Similarly we can consider the second-last element, third-last element, etc. using the exact same argument, we can show that the process stops.
Log in to reply
Great, that's 1 way. Find many many more!
Log in to reply
After I played around with a small list of numbers I noticed something interesting about the prime factorisation of the original sequence and the final sequence, that they are the same, but in the final sequence, they were sorted in an increasing order. (To see what I mean, try it out yourself!) This fact is rigorized by noticing that:
gcd(px,py)=pmin(x,y) and lcm(px,py)=pmax(x,y)
Then trivially this process would have to stop.
Just a thought: Maybe sum of the elements in the process might lead to the monovariant? (since I used product)
Log in to reply
Log in to reply
Log in to reply
I would use monovariance for the first problem, letting S equal the number of pairs (j,k) possible in the problem. Each successive iteration would reduce this number by 1, since aj and ak would be replaced by two numbers, the first of which does indeed divide the second, as gcd(aj,ak)∣lcm(aj,ak). Since there are a finite number of pairs, and each iteration reduces S by 1, eventually S will be zero, and the process could not be continued.
Log in to reply
Can you explain in more detail why the number of pairwise coprime integers is a monovariance? It is not immediately obvious to me.
If that works, that's 1 way. Find many many more!
Log in to reply
Oops. I misread the question and thought the two numbers were chosen such that they were coprime. I have changed the idea in my original comment.
Never mind. A proof that the sum is monovariant increasing and bounded from above is much harder than I thought.