Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Prove that for all $n\in\mathbb{N}, (n−1)^3+ n^3 \neq (n+1)^3$

Can someone give me a hint? I can only use properties of natural numbers... don't know how to write up a formal and correct proof for this one...

share|improve this question
    
yeah... sorry about that. –  Alex Lockwood Sep 11 '11 at 18:55
    
whoa i totally messed up with the edit then :) –  user13838 Sep 11 '11 at 18:56
    
sorry, i forgot to latexify it... :P –  Alex Lockwood Sep 11 '11 at 18:57
    
HINT Assume that the two sides are equal, and get a contradiction. (You may have to group together suitable terms.) –  Srivatsan Sep 11 '11 at 19:01
    
ok, thanks... i'll start with that! –  Alex Lockwood Sep 11 '11 at 19:02

6 Answers 6

up vote 4 down vote accepted

Expand and simplify $(n−1)^3+ n^3 -(n+1)^3$ and prove that it has no zero among the natural numbers.

This can be done as follows: $(n−1)^3+ n^3 -(n+1)^3 = n^3-6 n^2-2 = (n-6) n^2-2$. Now, if this is zero, then $(n-6) n^2=2$. Since the only square that divides $2$ is $1$, we must have $n=1$, which is not a solution of $(n-6) n^2=2$.

Or note that for $n\le 6$, we have $(n-6) n^2\le0$, and for $n\ge7$ we have $(n-6) n^2\ge49$.

share|improve this answer
    
yes... but how do you prove that there are no zeros? using the cubic equation formula seems way too messy... –  Alex Lockwood Sep 11 '11 at 19:10
    
@inradiobows, see my edited answer. –  lhf Sep 11 '11 at 19:16
    
ohhhhhhh... the "since the only square the divides 2 is 1" was key. that was the missing piece of the puzzle :) –  Alex Lockwood Sep 11 '11 at 19:20

Suppose the equality hold and just analyze the equation modulo n,

$(n−1)^3+ n^3 = (n+1)^3 \Rightarrow -1 \equiv 1 \pmod{n}$ or $n|2$, but 2 is not a solution. QED

share|improve this answer
1  
Note that this is essentially equivalent to applying the Rational Root Test to deduce that any integer root $\rm\:x = n\:$ of an integer coefficient polynomial $\rm\:f(x)\:$ necessarily divides the polynomial's constant term $\rm\:f(0)\:,\:$ i.e. if $\rm\ 0 = f(n)\:$ then mod $\rm\:n:\: \ 0 = f(n)\:\equiv\: f(0)\:,\:$ i.e. $\rm\:n\ |\ f(0)\:.\:$ See my answer. –  Bill Dubuque Sep 12 '11 at 3:56

Proof $1.\rm\ \ mod\ 4\:,\:$ the sequence of cubes repeats with period $4\:,\:$ namely $\rm\:\mathbb N^3 =\ 0,\ 1,\ 0,\:-1\:,\: \ldots\:.\:$
Four additions $\rm\:mod\ 4\:$ show that no element in the cycle is the sum of the prior two. $\ $ QED

Proof $2.\ $ $\rm\: (n-1)^3 + n^3 = (n+1)^3\:$ has no solutions $\rm\:mod\ 4\:$ so no integer solutions. This is easy to prove because cubes $\rm\:mod\ 4\:$ have very simple form: for even $\rm\:m = 2\:k\:$ we have $\rm\:m^3 = 8k^3\equiv\: 0\:,\:$ and odd $\rm\:m\:$ are $\equiv \pm1\:,\:$ so $\rm\:m^3\equiv m\:.\ $ Applying this to the equation, considered $\rm\:mod\ 4\:,\:$ we infer

$\begin{eqnarray} &\rm\ \: for&\rm even &\rm n\text{ the equation is}\ &\rm n\!\!-\!\!1\ &+\ \ 0\ &\equiv&\rm\ n\!\!+\!\!1&\rm\: \Rightarrow\ 2\: \equiv\: 0\ \ (mod\ 4)\ \ \ contradiction;\ \ \\ &\rm\ \: for&\rm\ odd &\rm n\ the\ equation\ is &\rm\ \ \ 0\ &+\rm\ \ n\ &\equiv&\rm\ \ \ \ 0&\rm \Rightarrow\ n\: \equiv\: 0\ \ (mod\ 4)\ \ \ contra\ n\ odd. \end{eqnarray}$ QED

NOTE $\ $ The advantage of the second proof is that the derived congruences for cubes are reusable and frequently useful; ditto for $\rm\ odd^2\:\equiv\:\{\pm1,\:\pm3\}^2\:\equiv\: 1\pmod{8}\:.$ Assuming known these ubiquitous congruences, the proof reduces to the above two displayed inferences - which can be derived in less than a minute of trivial mental arithmetic.

This technique of studying integer equations by studying their simpler (finite!) modular reductions in rings of integers modulo $\rm\:m\:$ is quite powerful. It is one of the fundamental problem-solving techniques in number theory and algebra.

Of course one could more simply apply the Rational Root Test to deduce that any integral root $\rm\:x = n\:$ of $\rm\:f(x) = (x+1)^3 - x^3 - (x-1)^3\:$ must divide $\rm\:f(0) = 2\:.\:$ The point of proceeding as above is to illustrate some general techniques which apply to less trivial Diophantine equations.

share|improve this answer

Another way to look at it: $(n-1)^3 + n^3$ is approximately $2n^3$, which is too large to be $(n+1)^3$, which is of the same order of magnitude as $n^3$ once $n$ gets to be a certain size.

More rigorously, if $n$ is big enough so that $(n+1)^3 < 2(n-1)^3 = (n-1)^3 + (n-1)^3$, you're also going to have $(n+1)^3 < (n-1)^3 + n^3$. But taking cube roots, $(n+1)^3 < 2(n-1)^3$ is equivalent to $$ (n + 1) < 2^{1 \over 3}(n-1)$$ Solving for $n$ you get $$n > {{2^{1 \over 3} + 1} \over {2^{1 \over 3} - 1}} = 8.6946...$$ So your equality can't hold if $n$ is at least $9$. You can check $n =1,2,...,8$ individually and verify the equality doesn't hold for them either. So it never holds.

share|improve this answer

HINT:

One can do this very quickly with induction for $n \geq 7$. Then there are only a finite number of cases to check.

share|improve this answer

This can be proved using various techniques.

One is to compare the left- and right-hand side modulo various small numbers and find one modulus where the two sides are never the same, and then prove it.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.