1

If $\frac{a(a^2+2)}{3}$ then $3\mid{a(a^2+2)}$.

By induction:

Lets define the set, $S=\left\{a\in N:a\geqslant1, 3\mid a(a^2+2) \right\}$

If $a=1$ then, $1\in S$

So we have to prove that if $k(k^2+2)=3m$ then $(k+1)((k+1)^2+2)=3n$ with $m,n\in Z$

If $k(k^2+2)=3m$ then,

$\begin{align*}k(k^2+2)+3(k^2+k+1)=&3m+3(k^2+k+1)\\=&3(m+k^2+k+1)\end{align*}$

Also,

$\begin{align*}k(k^2+2)+3(k^2+k+1)=&k^3+2k+3k^2+3k+3\\=&k^3+2k^2+k^2+2k+3k+3\\=&k^2(k+2)+k(k+2)+3(k+1)\\=&(k+2)(k^2+k)+3(k+1)\\ =&k(k+2)(k+1)+3(k+1)\\=&(k+1)(k(k+2)+3)\\ =&(k+1)(k^2+2k+1+2)\\ =&(k+1)((k+1)^2+2)\\ =&3(m+k^2+k+1) \end{align*}$

where $n=m+k^2+k+1$

Therefore,

$3\mid(k+1)((k+1)^2+2)$

Can I do it simplier using induction?

Grouper
  • 301
  • You haven't said what $S$ is. It's also quite hard to follow through. Perhaps include comments and numbering such as 'substituting ??? into (?), we get'. – Shuri2060 Aug 24 '17 at 22:43
  • What is meant by your first sentence? – MathTrain Aug 24 '17 at 22:45
  • You write $1\in S$. But what is $S$? – Shuri2060 Aug 24 '17 at 22:45
  • 1
    you can consider it mod 3. if a is 0,1,2 mod 3 what are the results ? if they all work then a can be any natural number and still work. –  Aug 24 '17 at 22:50
  • 2
    Can I do it simplier using induction? Not a lot simpler, but you can do it more easily without induction: $$ \frac{a(a^2+2\color{red}{+1-1})}{3} = \frac{3a}{3} + \frac{a(a^2-1)}{3}= a + \frac{(a-1),a,(a+1)}{3} $$ The numerator of the last fraction is the product of $,3,$ consecutive integers, so it is a multiple of $,3,$. – dxiv Aug 24 '17 at 22:52

6 Answers6

3

Yes, we can give a simpler and more conceptual inductive proof. Notice that

$\qquad\qquad a(a^2+2)\, =\, a(a^2\!-\!1 + 3)\, =\, \color{#0a0}{(a-1)a(a+1)} + 3a$

so it suffices to show that one of any $\rm\color{#0a0}{3\ consecutive\ integers}$ is divisible by $3$

This has a simple inductive proof. Note that shifting such a sequence by one simply replaces the old least element $\,\:\color{#C00}n\,$ by the new greatest element $\,\color{#C00}{n+3}$

$$ \begin{array}{} \:\color{#C00}n & n+1 & n+2 \\ \to & n+1 & n+2 & \color{#C00}{n+3} \end{array}$$

Since $ \: \color{#C00}n\equiv \color{#C00}{n+3} \pmod{\! 3},\,$ the shift does not change the set of remainders $\bmod 3$ of the elements. Thus the remainders remain the same as in the base case $ \ 0,1,2\: =\: $ all possible remainders mod $ \,3.\,$ Therefore the sequence has an element with remainder $\,0,\,$ i.e. an element divisible by $ \,3.$

Remark $ $ The same method works to show that a sequence of $d$ consecutive integers contains a multiple of $d.\,$ Alternatively this can be proved by using division with remainder (which has a natural proof by induction), which is closely connected to the above method.

Bill Dubuque
  • 272,048
2

Here is an alternative proof that doesn't use induction - just for fun!

Let $a \in \mathbb{Z}$. Observe that $a^3 - a = a(a^2 - 1) = (a-1)a(a+1)$ is the product of three consecutive integers, hence divisible by three. Adding $3a$ to the initial expression does not alter divisibility by $3$.

In particular, $a^3 - a + 3a = a^3 + 2a = a(a^2 + 2)$ is divisible by $3$, which means that it is still an integer after dividing by $3$. QED

(A ridiculous version of the above: $a(a^2+2)$ is the sum of $a^3 - a$ and $3a$, where the latter addend is a multiple of three by observation, and the former is divisible by three by Fermat's Little Theorem; so, their sum is divisible by three, and its quotient by three is an integer. QED)

1

I'd suggest writing the proof in a more linear fashion which is easier to read. Below is how I would approach the induction step.


Let $P(n)$ be the statement $3\mid n(n^2+2)$.

Proving $P(k)\implies P(k+1)$:

$$P(k)\implies 3\mid k(k^2+2)$$ $\implies$ $$3\mid k(k^2+2)+3(k^2+k+1)$$ $\implies$ $$3\mid k^3+3k^2+5k+3$$ $\implies$ $$3\mid(k+1)(k^2+2k+3)$$ $\implies$ $$3\mid(k+1)((k+1)^2+2)$$ $\implies$ $$P(k+1)$$

Shuri2060
  • 4,353
  • It has to do more with that I'm not yet fluent in latex. But I'll try and write it again following the way you do it – Grouper Aug 25 '17 at 01:09
  • 1
    @Mr.Lu You might find insightful to learn the idea behind the above induction step. Notice $${\rm if}\ \ 3\mid f(k)\ \ {\rm then}\ \ 3\mid f(k!+!1) \iff 3\mid f(k!+!1)-f(k)$$ So to prove the inductive step we need only verify $,3\mid f(k!+!1)-f(k) = 3(k^2+k+1),,$ which is clear. This is a special case of telescopic induction. – Bill Dubuque Aug 27 '17 at 18:01
0

If a is divisible by 3, then 3 divides a, and therefore a(a^2 + 2).

Otherwise, $a = 3k ± 1$, and 3 divides $(a^2 + 2)$ = $((3k ± 1)^2 + 2)$ = $((9k^2 ± 6k + 1) + 2)$ = $(9k^2 ± 6k + 3)$.

gnasher729
  • 10,113
0

Any number is congruent to $0$, $1$ or $2$ modulo $3$, and upon squaring you get any square is congruent to $0$, $1$ or $1$ modulo $3$. Now take an arbitrary number $a$. If $3$ divides $a$ then you're done, if not $a^2+2$ is, by the above, congruent to $0$ modulo $3$ since $1+2=3$, and so you're good, too.

Pedro
  • 122,002
  • 1
    This doesn't answer the question, i.e. "Can I do it simpler using induction" (my emphasis) since there is no (explicit) induction above. – Bill Dubuque Aug 24 '17 at 23:18
  • It does if you remove the last 'using induction'. OP can benefit from it, and learn how to write down proofs without filling up a whole page with symbols. – Pedro Aug 25 '17 at 08:09
0

Work backwards. It's far more natural.

$(k+1)((k+1)^2+2)=$

$(k+1)(k^2+2k+1+2)=$

$(k+1)(k^2+2)+(k+1)(2k+1)=$

$k (k^2+2)+(k^2+2)+(k+1)(2k+1)=$

$3m+(k^2+2)+(2k^2+3k+1)=$

$3m+3k^2+3k+3$.

It is sometimes easier to subtract results.

$(k+1)((k+1)^2+2)-k (k^2+2)=$

$k (k+1)^2+(k+1)^2+2k+2-k^3-2k=$

$k^3+2k^2+k +k^2+2k+1+2-k^3=$

$3k^2+3k+3$

is a multiple of 3.

But it's probably easiest not to use induction at all.

Let $a \equiv i \mod 3;i=0,\pm 1$

If $i=0$ then $3|a $ so $3|a (a^2+2) $

Otherwise $i=\pm 1$

So $a^2+2\equiv i^2+2\equiv 1+2\equiv 0\mod 3$

So $3|a^2+2$ so $3|a (a^2+2) $.

fleablood
  • 124,253
  • I'm not allowed to use mod yet. – Grouper Aug 25 '17 at 01:12
  • Then don't. Let $a = 3m +i$ where $i = 0, 1$ or $-1$. If $i = 0$ then $a(a^2+2) = 3m(a^2 + 2)$ is divisible by $3$. If $i = \pm 1$ then $a(a^2 + 2) = a((9m^2 \pm 6m +1)+ 2)=a(9m^2 \pm 6m + 3)$ is divisible by $3$. Those are the only three options. – fleablood Aug 25 '17 at 15:01