Find # Sequences for Increasing Sequence Problem

  • Thread starter Thread starter trollcast
  • Start date Start date
  • Tags Tags
    Sequences
AI Thread Summary
The discussion centers on determining how many sequences can yield a given increasing sequence as their average. The average sequence is defined by the equation a[i] = 0.5(s[i] + s[i+1]), where s is the original sequence. Participants explore the relationship between the differences of consecutive terms in the sequence and how adjusting these can create valid sequences. Specific examples illustrate that while some sequences, like 1, 2, 3, 4, have multiple valid predecessors, others, such as 1, 2, 8, 9, do not yield any valid sequences. The conversation emphasizes the need for the resulting sequences to remain increasing or non-decreasing.
trollcast
Gold Member
Messages
282
Reaction score
13

Homework Statement


For an increasing sequence of numbers, how many other sequences could this be the average sequence of.

Homework Equations



Where the average sequence, a = 0.5( s + s[i+1] )

The Attempt at a Solution



If there's n terms in the original sequence.
The number of differences between consecutive terms is (n - 1)
Find all these, (n-1) and find the lowest difference.
Then this lowest difference + 1 is your answer?

eg.
s = 1, 3, 6, 10, 12
s[i+1] - s = 2, 3, 4, 2

The lowest difference here is 2 so there's 2 possible sequences for which s is the average of?
 
Physics news on Phys.org
trollcast said:

Homework Statement


For an increasing sequence of numbers, how many other sequences could this be the average sequence of.

Homework Equations



Where the average sequence, a = 0.5( s + s[i+1] )

The Attempt at a Solution



If there's n terms in the original sequence.
The number of differences between consecutive terms is (n - 1)
Find all these, (n-1) and find the lowest difference.
Then this lowest difference + 1 is your answer?

eg.
s = 1, 3, 6, 10, 12
s[i+1] - s = 2, 3, 4, 2

The lowest difference here is 2 so there's 2 possible sequences for which s is the average of?


Hey trollcast! :smile:

Suppose we start with the simplest possible sequence we can think of: 1,2,3,4,...

Now we'd look for another sequence that has the same averages.
To get the same average, if we increase a specific number by some amount ε > 0, the next number must be decreased by that same amount ε.
So we'd add an alternating sequence ##(-1)^i ε##.
Since the resulting sequence still has to be increasing, that ε must be less than (1/2).

So we'd start with the sequence specified by ##s_i = i## and we'd end up with the sequence specified by ##s_i' = i + (-1)^i ε## with ##0 < ε < \frac 1 2 ##.

How many of those ε's are there?
 
I like Serena said:
Hey trollcast! :smile:

Suppose we start with the simplest possible sequence we can think of: 1,2,3,4,...

Now we'd look for another sequence that has the same averages.
To get the same average, if we increase a specific number by some amount ε > 0, the next number must be decreased by that same amount ε.
So we'd add an alternating sequence ##(-1)^i ε##.
Since the resulting sequence still has to be increasing, that ε must be less than (1/2).

So we'd start with the sequence specified by ##s_i = i## and we'd end up with the sequence specified by ##s_i' = i + (-1)^i ε## with ##0 < ε < \frac 1 2 ##.

How many of those ε's are there?

Would there be 0 if the sequence has to be integers?
 
trollcast said:
Would there be 0 if the sequence has to be integers?

Yes if we're talking about the sequence 1,2,3,...

But for another sequence like 0,10,20,30,40,...
we can find 1,9,21,29,41,...
or 2,8,22,28,42,...
or ...
 
I like Serena said:
for another sequence like 0,10,20,30,40,...
we can find 1,9,21,29,41,...
or 2,8,22,28,42,...
or ...
Maybe I'm misreading the OP. If a = 0,10,20,30,40,.. is the given sequence, I thought we were looking for other sequences s s.t. a = 0.5( s + s[i+1] ). 1,9,21,29,41,... does not appear to be a solution.
If we start with an arbitrary s[0] then s[i+1] = 2a - s would appear to generate the rest uniquely. Perhaps it is also required that s is increasing (or maybe non-decreasing)?
 
haruspex said:
Maybe I'm misreading the OP. If a = 0,10,20,30,40,.. is the given sequence, I thought we were looking for other sequences s s.t. a = 0.5( s + s[i+1] ). 1,9,21,29,41,... does not appear to be a solution.
If we start with an arbitrary s[0] then s[i+1] = 2a - s would appear to generate the rest uniquely. Perhaps it is also required that s is increasing (or maybe non-decreasing)?


Yeah that's how the example worked, the new sequence must be non decreasing as well.
 
Actually I looked at this again and for the easiest one, ie. the average sequence 1, 2, 3, 4, ... , there's 2 possible sequences that could have came from:

0, 2, 2, 4, 4, 6, 6, ...

or

1, 1, 3, 3, 5, 5, ...

I've sat and mucked around with various other sequences and found that something like:

1, 2, 8, 9 has no sequences for which it could be the average but I can't figure out a way to work this out without going through and working out the possible values for each average sequence?
 
Back
Top