Skipping sums

Algebra Level 3

If a + a 2 + a 3 + a 4 + a 5 + a + a^2 + a^3 + a^4 + a^5 + \cdots is a positive number, then which of the following is larger,

a + a 3 + a 5 + a 7 + or a 2 + a 4 + a 6 + a 8 + ? a+a^3+a^5+a^7+\cdots\quad \text{or}\quad a^2+a^4+a^6+a^8+\cdots\, ?

a + a 3 + a 5 + a 7 + a + a^3 + a^5 + a^7 + \cdots a 2 + a 4 + a 6 + a 8 + a^2 + a^4 + a^6 + a^8 + \cdots They are equal

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

11 solutions

Marta Reece
Jun 4, 2017

For the original geometric series to be convergent, it has to have a common ratio r = a < 1 |r|=|a|<1 .

This means that each term in the series is smaller in absolute value than the preceding term.

For the expression

( a + a 2 ) + ( a 3 + a 4 ) + ( a 5 + a 6 ) + = ( a + a 2 ) + a 2 ( a + a 2 ) + a 4 ( a + a 2 ) + = ( a + a 2 ) ( 1 + a 2 + a 4 + ) (a+a^2)+(a^3+a^4)+(a^5+a^6)+\cdots=(a+a^2)+a^2(a+a^2)+a^4(a+a^2)+\cdots =(a+a^2)(1+a^2+a^4+\cdots)

to be positive, a a , which is larger in absolute value than a 2 a^2 , has to be positive.

The two series in question

a + a 3 + a 5 + a 7 + a + a^3 + a^5 + a^7 + \cdots

a 2 + a 4 + a 6 + a 8 + a^2 + a^4 + a^6 + a^8 + \cdots

can now be compared term by term. All of the terms in both series are positive, and each term of the first series is larger than the corresponding term of the second series.

The sum of the first series is therefore larger than the sum of the second series.

It was not stated that the series converged, merely that a was a positive number. Therefore for most values of a, a>=1, the series diverge, and are therefore equal in value (infinity). An unusually poorly worded question.

Paul Vincent - 3 years, 12 months ago

Log in to reply

If an infinite sum is a positive number, the series is convergent. Infinity is not a number, positive or otherwise.

Marta Reece - 3 years, 12 months ago

My apologies - I misread the question as saying that a was a positive number, whereas of course it states that the SERIES is a positive number, which - as others have stated - implies convergence. My silly mistake.

Paul Vincent - 3 years, 12 months ago

A quibble that is not a problem here "is a positive number" can be reasonably taken to mean "converges" here. At least I so took it without offense.

Robert DeLisle - 3 years, 12 months ago

@Marta Reece - If infinite is neither positive or negative, then why is it marked on the axis' extremities with + and - before them?

Victor Shirosaki - 3 years, 12 months ago

Log in to reply

Infinity can be positive or negative, but it still is not a number.

Marta Reece - 3 years, 12 months ago

To me, as mathematical non-expert, "a" must be a positive number, and it is not made clear from the wording that a has to be between zero and 1. Not to me. If you want to make that clear, then you should have spelled it out. This is supposed to be a math puzzle, not a "legalistic masterpiece contract" with "invisible fine print".

So, when I assume "a" to be 0.25, then the first series is always bigger. When I assume "a" to be 1, the series are equal, and when I assume "a" to be bigger than 1, then the second series is always bigger. I, being simple-minded, of course consider a "positive infinity" to be still "a positive number". (And a negative infinity is still a negative number, likewise. Just because I don't know HOW big it is, I can still assign positive and negative values. What law of math can forbid me that? ) So depending on the input, and the non-existence of constraints, I can come up with three answers: Bigger, equal and lower. Real life comparisons can only be made when one makes a specific point in time evaluation. (I know, that's why they say "put three accountants in a room, and you get six different answers".)

I calculated out the value of a compared to a^2, a+a^3 compared to a^2+a^4, and so forth, because I simply and idealistically assumed an invisible and unprovable "fairness in the mathematical universe", such that each of these two series each would have to have the same number of terms to be comparable.
If I am assuming too much, then the writer of this puzzle likewise assumed too much.

My math teacher in high-school always was annoyed when I calculated things out, in order to get a feel for where I was getting at with the result. I guess that's why I ended up as tax guy, because the answers have to make sense, to me, the client and the IRS. Infinite tax bills or infinite refunds are just never the right answer.

Thank you for your patience!

Not really Dontyathink - 3 years, 12 months ago

How is each term of the first series larger than each term of the second series?

Aditya Shukla - 3 years, 12 months ago

Log in to reply

If a a is a positive number less than 1 1 , multiplying by a a will make any positive number smaller. Each term by term comparison will compare a n a^n to a n + 1 = a n × a a^{n+1}=a^n\times a . The second number is the same as the first, except for being multiplied by a a . It is therefore smaller.

Marta Reece - 3 years, 12 months ago

I failed to see that the dots indicated an infinite series, which therefore had to have |a|<1 for it to converge. In a broader sense of the terms, the dots could have meant that this was just a simple sum of finite length up to the nth term, which in turn would have meant that for every a such that |a|>1, the even powers sum diverges faster than the other one. which means that in all probability the answer was the other one. But then again, these are maths, so there cannot be a 'mostly true answer' for such problems therefore I should have gotten the point that the sums were infinite

Alessandro Jeanteur - 3 years, 12 months ago

Log in to reply

Continued dots is a standard maths notation for a series containing an infinite number of elements

Stephen Smith - 3 years, 12 months ago

No. You have neglected to consider solutions where a<1. All even powered terms will be positive, but odd powered terms of negative values of a will be negative. There the answer is the even powered series is larger.

Donald Zacherl - 3 years, 12 months ago

Log in to reply

I have handled the possibility. I have factored the entire sum into a product of an obviously positive one - one make of the number 1 and only even powers of a a - times a + a 2 a+a^2 . If a < 1 |a|<1 , then a 2 < a |a^2|<|a| . So for a + a 2 a+a^2 to be positive, a a has to be positive.

Marta Reece - 3 years, 12 months ago

Since "depends on sign of a" is not one of the choices, it is sufficient to work it out with a positive value of a, 0 < a < 1 (so bounded by the implied convergence "is a positive number" to eliminate "equal". If the other case did not work out, THEN it is a poorly worded problem. The quick and dirty tactical approach.

Robert DeLisle - 3 years, 12 months ago
Uros Stojkovic
Jun 12, 2017

First thing you should conclude is that 0 < a < 1 0< a< 1 . Why? Well, first sentence says that the sum of a + a 2 + a 3 + a 4 + . . . a + a^{2} + a^{3} + a^{4} + ... is a positive number, and that means that this infinite series converges to some positive value. If a > 1 a> 1 , then the sum would go to the infinity. Also, a a can't be negative. Here is the math:

a + a 2 + a 3 + a 4 + a 5 + . . . = a 1 a > 0 a + a^{2} + a^{3} + a^{4} + a^{5} + ... = \frac{a}{1-a}> 0

( a > 0 1 a > 0 ) \Rightarrow (a> 0 \wedge 1-a > 0) or ( a < 0 1 a < 0 ) (a < 0 \wedge 1-a < 0) - both numerator and denominator must have the same sign.

Solution of the system a > 0 1 a > 0 a> 0 \wedge 1-a > 0 is set of numbers 0 < a < 1 0< a< 1 , and the system a < 0 1 a < 0 a < 0 \wedge 1-a < 0 has no solutions.

Hence 0 < a < 1 0< a< 1 .

Now, once we cleared that, we can move on with our problem:

a + a 3 + a 5 + a 7 + . . . = a 1 a 2 a + a^{3} + a^{5} + a^{7} + ...= \frac{a}{1-a^{2}} and

a 2 + a 4 + a 6 + . . . = a 2 1 a 2 a^{2} + a^{4} + a^{6} + ...= \frac{a^{2}}{1-a^{2}}

We can notice that they have the same denominator, so the one with greater numerator will be greater. Since 0 < a < 1 0< a< 1 , then a > a 2 a> a^{2} .

Hence, a + a 3 + a 5 + a 7 + . . . > a 2 + a 4 + a 6 + . . . a + a^{3} + a^{5} + a^{7} + ...> a^{2} + a^{4} + a^{6} + ... .

When I attempted this I didn't realise that "A positive number" excludes infinity.

Justin Roughley - 3 years, 12 months ago

Log in to reply

Yeah, that's a common mistake, but, one must realize that infinity is not a number! It's just a concept of unbounded limit, when you have i = 0 f ( i ) = \sum_{i=0}^{\infty } f\left ( i \right )= \infty that doesn't mean that this sum converges to some value (infinity as a number), but that it grows without bound never reaching any final value. I hope you understand what I'm saying. :)

Uros Stojkovic - 3 years, 12 months ago

as they are BOTH infinite series, they are (for all practical purposes) equal...

Bob Vandenbosch - 3 years, 12 months ago

Log in to reply

It's true that they are both infinite series, but that doesn't mean that they are the same. The crucial thing is to see that 0 < a < 1 0< a< 1 . If you aren't still convinced, use the calculator, plug in value of a a such that 0 < a < 1 0< a< 1 and then compare the sums of e.g. first 3 terms.

Uros Stojkovic - 3 years, 12 months ago

I believe you typed a-1 instead of 1-a 4 times in your solution when showing that 0 < a < 1

Jacob Huebner - 3 years, 12 months ago

Log in to reply

You are right, I've just corrected it. Thanks for the note!

Uros Stojkovic - 3 years, 12 months ago
Tom Engelsman
Jun 4, 2017

The original infinite geometric series sums to:

a + a 2 + a 3 + a 4 + a 5 + . . . = a 1 a = N a = N N + 1 a + a^2 + a^3 + a^4 + a^5 + ... = \frac{a}{1-a} = N \Rightarrow a = \frac{N}{N+1} , where N R + . N \in \mathbb{R^{+}}.

Let S o d d S_{odd} and S e v e n S_{even} be the sum of the odd and the even exponent terms respectively. These yield:

S o d d = a + a 3 + a 5 + . . . = 1 a a 2 1 a 2 S_{odd} = a + a^3 + a^5 + ... = \frac{1}{a} \cdot \frac{a^2}{1 - a^2} and S e v e n = a 2 + a 4 + a 6 + . . . = a 2 1 a 2 . S_{even} = a^2 + a^4 + a^6 + ... = \frac{a^2}{1-a^2}.

Since S o d d = 1 a S e v e n = N + 1 N S e v e n = ( 1 + 1 N ) S e v e n , S_{odd} = \frac{1}{a} \cdot S_{even} = \frac{N+1}{N} \cdot S_{even} = (1 + \frac{1}{N}) \cdot S_{even}, we conclude S o d d > S e v e n . \boxed{S_{odd} > S_{even}}.

Who said it has to be convergent? If it isnt then the answer is wrong.

Rodion Zaytsev - 3 years, 12 months ago

Rodion, By assumption from the problem statement: "If [series] is a positive number..."

J Yoest - 3 years, 12 months ago

I believe it is a tricky question. If the number is 0<a<1 yes then the a+a3.. converges and the value is indeed larger than a2+a4 But if a=1 then they are the same as S odd - S even =0

Rodolfo Ruiz Huidobro - 3 years, 12 months ago
Kelvin Hong
Jun 12, 2017

If we want a + a 2 + a 3 + a 4 + . . . a+a^2+a^3+a^4+... to be a positive number but no diverge to infinity, we must assume 0 < a < 1 0<a<1

So,

S e v e n = a S o d d S_{even}=aS_{odd}

S e v e n < S o d d \boxed{S_{even}<S_{odd}}

using a=1 they are equal and using a=2 the other awnser is correct... a is not given as 0<a<1 so there is no reason to suppose this

stephen denboer - 3 years, 12 months ago

Log in to reply

It is a infinity geometric expression, if the result is a positive number that means converge, so must have 0<a<1.

Kelvin Hong - 3 years, 12 months ago
Jaka Kranjc
Jun 12, 2017

Since the geometric series' sum is a number, it has to converge and since it is a positive number, 0 < a < 1 0 < a < 1 .

Looking at the two sums in question, we can rewrite the second as a product of the first: a ( a + a 3 + a 5 + ) a(a+a^3+a^5+\ldots) . Since a a is smaller than 1, the second sum must be smaller than the first.

Why must 0<a<1 be true? We only know that -1<a<1 is true (because the GP converges), so why can't -1<a<0 be true as well?

Pi Han Goh - 3 years, 11 months ago

Log in to reply

The sum can't be positive if a is negative.

Jaka Kranjc - 3 years, 11 months ago

Log in to reply

Why is that the case?

Pi Han Goh - 3 years, 11 months ago

Log in to reply

@Pi Han Goh The leading term will dominate the sum. If you want a proper proof of this, Marta and Uroš have two variants included.

Jaka Kranjc - 3 years, 11 months ago
Lorenzo Calogero
Jun 15, 2017

Let A = a 2 + a 4 + a 6 + . . . = a ( a + a 3 + a 5 + . . . ) A=a^2+a^4+a^6+...=a(a+a^3+a^5+...) and B = a + a 3 + a 5 + . . . B=a+a^3+a^5+...

A B = a \frac{A}{B}=a . Since a + a 2 + a 3 + . . . > 0 a+a^2+a^3+...>0 , 0 < a < 1 0<a<1 ; therefore, 0 < A B < 1 B > A 0<\frac{A}{B}<1 \Rightarrow \boxed{B>A}

Yes, A/B > 0 is a positive number, but is it possible that A<0 and B<0 as well? If so, then maybe it's still possible that A<B, no?

Pi Han Goh - 3 years, 11 months ago
Massimo 22
Jun 9, 2017

Wrote a bit of python code:

    x = 2

    b = 0

    c = 1


    for i in range(1,10):
        a = x ** b
        print a
        c += 2

    for i in range(1,10):
        a = x ** c
        print a
        c += 2

Plugging in a few values for x, we can see that the system with the odd exponents is always larger.

This is wrong. You have only compare the values of 2 finite sums, whereas the question asked you to compare the values of 2 infinite sums.

Pi Han Goh - 4 years ago
Richard Desper
Jun 16, 2017

Use the formula i = 0 r i = 1 1 r \sum_{i=0}^{\infty} r^{i} = \frac{1}{1-r} , which is valid for r < 1 |r| < 1 .

Since the original series converges, its sum is S 1 = a 1 a S_1 = \frac{a}{1-a} . We are also told that it is positive, which implies that a > 0 a > 0 .

We are next asked to compare two sums, each of which is a geometric series with ratio a 2 a^2 . Since a < 1 |a| < 1 it follows that a 2 < 1 |a^2| < 1 , so we know both of these series must also converge. Their sums are S 2 = a 1 a 2 S_{2} = \frac{a}{1-a^2} and S 3 = a 2 1 a 2 S_3 = \frac{a^2}{1-a^2} . Thus we see that S 3 = a S 2 S_3 = a * S_2 . Since 0 < a < 1 0 < a < 1 , this implies that S 3 < S 2 S_3 < S_2 .

How do you know that 0<a<1 is true? Or how did you rule out -1<a<0?

Pi Han Goh - 3 years, 11 months ago
Peter Hauge
Jun 15, 2017

For the original series to converge to a positive number, 0<a<1. You get the second series by multiplying the first by a, so the second series converges to a smaller number.

How did you immediately know that 0<a<1 is true? We know that -1<a<1 forces the original series to converge, but how do you know that -1<a<0 cannot be fulfilled either?

Pi Han Goh - 3 years, 11 months ago
Ali Ouahassoune
Jun 13, 2017

let U n = i = 0 n a 2 i + 1 U_{n}= \sum_{i=0}^{n} a^{2i+1} and V n = i = 1 n a 2 i V_{n}= \sum_{i=1}^{n} a^{2i} with n \in N U n V n = i = 0 n a 2 i + 1 i = 1 n a 2 i = i = 0 n a 2 i + 1 i = 0 n a 2 i + 1 = i = 0 n a 2 i + 1 a 2 i + 1 = i = 0 n a 2 i ( a 1 ) + 1 U_{n} - V_{n} = \sum_{i=0}^{n} a^{2i+1} - \sum_{i=1}^{n} a^{2i} = \sum_{i=0}^{n} a^{2i+1} - \sum_{i=0}^{n} a^{2i} + 1 = \sum_{i=0}^{n} a^{2i+1} - a^{2i} + 1 = \sum_{i=0}^{n} a^{2i}( a - 1 ) + 1 Therefore: If a > 1 a > 1 then U n V n U_{n} - V{n} is positive. so U n > V n U_{n} > V_{n}

But if you substitute any value of a a such that a > 1 a>1 , then the original series diverges to infinity, so your solution is invalid.

Pi Han Goh - 3 years, 11 months ago
Aneesh Saripalli
Jun 13, 2017

For the value of the infinite series to be a real number (i.e. converge) the r < 1 **|r| < 1** . Because both series (one of even exponents, one of odd exponents) have real values, it's fair to assume they have a real number sum. Hence a 2 + a 4 + a 6 . . . a^2 + a^4 + a^6... = ( a + a 3 + a 5 . . . ) a (a + a^3 + a^5...) * a . Because we know a < 1 |a| < 1 and that a proper (positive) fraction of a (positive) value has to be less than that same value, we conclude that ( a + a 3 + a 5 . . . ) > ( a 2 + a 4 + a 6 . . . ) (a + a^3 + a^5...) > (a^2 + a^4 + a^6...)

Because both series (one of even exponents, one of odd exponents) have real values, it's fair to assume they have a real number sum.

It's a good (and correct) assumption, but for rigor, can you prove that it's true?

Pi Han Goh - 3 years, 11 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...