2

Prove that for all positive integers n, $20^{2n} + 16^{2n} −3^{2n} −1$ is a multiple of 323

I tried to prove by induction, no luck

Step 1: Base Case (n = 1)

$20^{2(1)} + 16^{2(1)} - 3^{2(1)} - 1$

$= 646$

The expression is divisible by 323 when n = 1

Step 2: Inductive Hypothesis

Assume that for some positive integer k, the expression $20^{2k} + 16^{2k} - 3^{2k} - 1$ is divisible by 323.

Step 3: Inductive Step

Consider the expression for k+1:

$20^{2(k+1)} + 16^{2(k+1)} - 3^{2(k+1)} - 1$

We can rewrite it as follows:

$20^{2(k+1)} + 16^{2(k+1)} - 3^{2(k+1)} - 1$ $= (20^2)^{k+1} + (16^2)^{k+1} - (3^2)^{k+1} - 1$ $= (400)^{k+1} + (256)^{k+1} - (9)^{k+1} - 1$

We assumed that:

$20^{2k} + 16^{2k} - 3^{2k} - 1$ is divisible by 323.

So, we can write it as:

$20^{2k} + 16^{2k} - 3^{2k} - 1 = 323m$ for some integer m.

$400^{k} + 256^{k} - 9^{k} - 1 = 323m$

Now, we can substitute this into our expression for k+1:

$(400)^{k+1} + (256)^{k+1} - (9)^{k+1} - 1$

Bill Dubuque
  • 272,048
Mzq
  • 252
  • 1
    What have you tried? show your attempts. – Lion Heart Sep 04 '23 at 09:33
  • 4
    Have you considered breaking it down into sub-problems? $323$ has a prime factorization of $17\times 19$, so if $323|x$, then $17|x$ and $19|x$. Those numbers are much easier to work with; furthermore, you might find it profitable to write $20, 16$ and $3$ as $17k+r$ and $19k+r$ for some integers $k$ and $r$ and making use of the binomial theorem... no messy induction argument required. – H. sapiens rex Sep 04 '23 at 09:42
  • First reduce your expression modulo 19 and check it always divides 19. Second reduce it modulo 17 and check it always divides 17. – quarague Sep 04 '23 at 09:45
  • Exploit innate $\rm\color{#c00}{symmetry}$ as in linked dupe. $\tag{}$ $\qquad\phantom{\Rightarrow}\ \ { 20^2,\ \ \ \ 16^2}\ \equiv, {3^2,\ \ \ 1^2}\ \ \ \ {\rm mod},\ 17,19,\ \ $
    $\tag
    {}$ $\qquad\Rightarrow\ {20^{2k},\ \ 16^{2k}} \equiv {3^{2k},\ ,1^{2k}}\ \ \ {\rm mod},\ 17,19,\ \ $ by the Congruence Power Rule
    $\tag*{}$ $\qquad\Rightarrow\ \ 20^{2k} + 16^{2k}\ \ \equiv \ 3^{2k}+ 1^{2k}\ \ \ \ {\rm mod},\ 17,19,\ \ $ so also $,{\rm mod}\ 323 = {\rm lcm}(17,19)$ by CCRT.$\ \ $
    – Bill Dubuque Sep 04 '23 at 16:59
  • Please search for answers before posting questions, e.g. here searching on "323 20" quickly finds many dupes. See the sidebar "Linked" questions in the dupes for many more worked examples. – Bill Dubuque Sep 04 '23 at 19:35

2 Answers2

3

Expanding my comment into an answer:

$323$, being much larger than any of the numbers in the given expression, is unwieldy to work with. It is better to make use of its prime factorization — which is $17\times19$ — and the fact that a number is divisible by $323$ if and only if it is simultaneously divisible by $17$ and $19$. Here's how that proof goes.

Working with $17$ first, rewrite $20$ as $17+3$ and $16$ as $17-1$. It would be awkward to rewrite $3$ in this way however, so we shall leave it alone. Next, consider the binomial expansion of $(17+3)^{2n}$ and its reduction modulo $17$. It is not too hard to see that all but one of the terms in the expansion contain at least one factor of $17$; thus, the remainder modulo $17$ is simply $3^{2n}$. That cancels out with the $-3^{2n}$, conveniently enough. Next, consider the binomial expansion of $(17-1)^{2n}$. By exactly the same argument, the remainder of this expansion modulo $17$ is nothing but $(-1)^{2n}$. Since $(-1)^m=1$ for all even numbers $m$, and $2n$ is even for all values of $n$, we have that $(-1)^{2n}=1$ for all values of $n$. This cancels out with the $-1$ in the given expression; thus the entire sum is congruent to $0 \mod(17)$, proving that it is divisible by $17$ for all values of $n$.

The other case proceeds in exactly the same way. Rewrite $20$ as $19+1$ and $16$ as $19-3$, then consider the reduction of $(19+1)^{2n}$ and $(19-3)^{2n}$ modulo $19$: all but one term of $(19+1)^{2n}$ contains at least one factor of $19$, leaving $1^{2n}=1$ as the remainder; this cancels with the $-1$ in the given expression. Likewise, all but one term of $(19-3)^{2n}$ contains a factor of $19$, leaving $(-3)^{2n}$ as the remainder; furthermore, just as $(-1)^{2n} = 1^{2n}$ for all values of $n$, $(-3)^{2n} = 3^{2n}$ for all values of $n$, which means that it cancels with the $-3^{2n}$ in the original expression. Thus the original expression is congruent to $0 \mod(19)$ for all values of $n$.

We have successfully proved that for all values of $n$, $20^{2n}+16^{2n}-3^{2n}-1$ is congruent to $0$ modulo $17$ and $19$, which by definition means that it is congruent to $0$ modulo $323$. Thus the claim is proved. All without induction!

  • Please strive not to post more (dupe) answers to dupes of FAQs, cf. recent site policy announcement here. Note that this dupe is very easy to find, e.g. search on "323 20". – Bill Dubuque Sep 04 '23 at 19:28
1

Assuming you know congruence theory this exercise is pretty straightforward, i.e:

From $20=17+3=19+1$ and $16=17-1=19-3$ you get

$\begin{cases} 20^{2n}+16^{2n}-3^{2n}-1\equiv (+3)^{2n}+(-1)^{2n}-3^{2n}-1\equiv 0\pmod{17}\\ 20^{2n}+16^{2n}-3^{2n}-1\equiv (+1)^{2n}+(-3)^{2n}-3^{2n}-1\equiv 0\pmod{19} \end{cases}$

So the expression is simultaneously divisible by $17$ and $19$ and consequently by $323$.

Now if you have never heard of modulo, just consider (for $n\ge 1$) the binomial expansion of

$\begin{align}(m+a)^n&=a^n+\binom{n}{1}a^{n-1}m+\binom{n}{2}a^{n-2}m^2+\cdots+m^n\\&=a^n+m\times(\cdots)\end{align}$

You can see that in the division by $m$ the terms $(m+a)^n$ and $a^n$ have the same remainder (i.e. their difference is a multiple of $m$); we say they are equivalent relatively to the division by $m$.

Considering $m=17$ and $a=3$ then $20^{2n}$ and $3^{2n}$ are equivalent and get cancelled by the trailing term $-3^{2n}$ in the expression.

Similarly with $a=-1$, it get cancelled thanks to the even exponent and the trailing term $-1$.

And we can say the same for $m=19$.


If you really prefer induction, then if it advisable to take advantage of the facts seen above.

$\begin{align}20^{2n+2}+16^{2n+2}-3^{2n+2}-1 &=20^{2n}(17+3)^2+16^{2n}(17-1)^2-9\times 3^{2n}-1\\ &=17\times(\cdots)+9\times 20^{2n}+16^{2n}-9\times 3^{2n}-1\\ &=17\times(\cdots)+9\times (\underbrace{20^{2n}-3^{2n}}_\text{divisible by 17}) + (\underbrace{16^{2n}-1}_\text{divisible by 17})\\ \end{align}$

But you see that you have to carry out a double hypothesis of divisibility by $17$ and not just a single one under the form of a sum which makes it more intractable.

zwim
  • 28,563
  • Please strive not to post more (dupe) answers to dupes of FAQs, cf. recent site policy announcement here. Note that this dupe is very easy to find, e.g. search on "323 20". – Bill Dubuque Sep 04 '23 at 19:28