1

The given problem is the following one.

Find the maximum natural number which is unable to be represented as $17\times x +23\times y \quad x,y\in \mathbb{N}$ .

I've been solving it till following formulas with looking the solution of the problem.

$1\leqq i \leqq 22 \quad 17 \times i \equiv r_i(mod 23)$

$17\times i\not\equiv0(mod23)$

$n(r)=22$

Therefore we can assume the problem's formula as $(1\leqq x\leqq23 \land y,q\in\mathbb{N})\quad 17\times x+23\times y =23\times q+r\prime \quad\dots(1)$ .

$\uparrow $Because if x is greater than 23,then $17\times x$ is represented as $17\times(23\times t+u)\quad (t\in \mathbb{N,u \in\{1,2,...,23\}})$

$=17\times (23\times t)+17\times u ...(2)$

and $1$st term of $(2)$ is able to be absorbed in $2$nd term of$(1)$.

And my problem begins from here.

The solution says that it is possible to be represented the all natural numbers which are greater than $17\times 23$ with $(1)$and it is impossible to be represented $17\times23$,so the answer is $17\times 23$.

Of course I can deduce that $17\times 23 $ is unable to be represented as a form of $(1)$ but I can't understand that"it is possible to be represented the all natural numbers which are greater than $17\times 23$".

How could this proven?.

  • 2
    Apply the Chicken McNugget Theorem: https://artofproblemsolving.com/wiki/index.php/Chicken_McNugget_Theorem (link contains proofs as well) – zhuli Jan 03 '20 at 08:38
  • 2
    The number you are looking for is called the Frobenius number. – N. F. Taussig Jan 03 '20 at 09:45
  • 1
    @[email protected] I get $17\times 23 -(17+23)$ as a answer when I use Chicken McNugget Theorem and this answer is incorrect to the solution. –  Jan 03 '20 at 17:20
  • So it the terminology of "Chicken McNugget Theorem" slowly becoming more well known then "Frobenius Coin Problem"? .... sigh... stupid modern world...... My grandfather would turn is his grave. (And once he turned in his grave he'd exchange it for a convertible and .....) – fleablood Jan 03 '20 at 17:38
  • I understand how you feel but ultimately I remembered "chicken McNugget" and could not remember "Frobenius coin". As much as I wish it wasn't so, branding is a real phenomenon. – zhuli Jan 03 '20 at 17:56
  • Okay.. I'm sorry but that cited link says: "There are many stories surrounding the origin of the Chicken McNugget theorem. However, the most popular by far remains that of the Chicken McNugget" but then it says "has also been called the Frobenius Coin Problem or the Frobenius Problem, after German mathematician Ferdinand Frobenius inquired about the largest amount of currency that could not have been made with certain types of coins". As that is the EXACT same question and predates McDonalds by 100 years, there is NO WAY we can say the theory "originated" as Chicken McNugget. !!GRUMPH!! – fleablood Jan 03 '20 at 19:32
  • Nobody here claimed the theory originated as "chicken McNugget". All I said was it was easier to remember personally. – zhuli Jan 03 '20 at 21:09
  • @N.F.Taussig Actually it is the modified Frobenius number since here we seek positive (vs. nonnegative) solutions. – Bill Dubuque Jan 04 '20 at 02:21
  • fyi: beware that the 2nd proof in the answer you accepted has errors (see my comments there). Hopefully it did not lead to any confusion. The simplest way to do this is to use arithmetic (vs. induction) as in the method in my answer. – Bill Dubuque Jan 05 '20 at 23:32

1 Answers1

1

$ 17x\!+\!23y = \color{#0a0}{n > 17(23)}\,$ is solvable for $\,x,y\ge 1,\,$ by mod $23\!:\, $ there's $\,x\equiv 17^{-1}n,\ 1\le \color{#c00}{x \le 23},\,$ so $ 17x \equiv n,\,$ so $\,17 x + 23 y = n,\,$ for $\, y\in\Bbb Z,\,$ and $\,y>0\,$ by $\,17\color{#c00}x \le 17(\color{#c00}{23})\color{#0a0}{< n}$

$ 17x\!+\!23y\: {\bf\color{#c00}=}\: 17(23)\,$ is $\rm\color{#c00}{unsolvable}$: $\, 17\mid 23y\Rightarrow 17\mid y\ $ so $\ x\!+\!23\:\!{\large \frac{y}{17}}\! = 23\,$ contra $\,x,{\large \frac{y}{17}} \ge 1$

We used: if $\,\gcd(a,b)=1\,$ then $\,a^{−1}$ exists $\!\bmod b\,$ (e.g. via Bezout gcd identity) and, furthermore $\,a\mid bc\Rightarrow a∣c\,$ (by Euclid's Lemma), for $\,a,b=17,23\,$ above. Since we used only coprimality of $\,17,23,\,$ the proof works for any coprime $\,a,b > 1,\,$ e.g. see here, which has a more geometric proof, along with citations on this well-known Frobenius Coin Problem.

Remark $\, $ A unit shift translates the above to permit $\,x,y = 0,\,$ namely $\ \ \ \ \begin{align} &\ \ \ \ \, 17\,x^{\phantom{|^|}} \ \,+\ \ \ \ 23\,y\, \ \ \ =\ \ \ \ n\qquad\quad\ \ \,{\rm for}\ \ x,y \ge 0\\[.2em] \iff\ &17(x\!+\!1) + 23(y\!+\!1) =\, n\!+\!17\!+\!23\,\ \ {\rm for}\ \ x\!+\!1,y\!+\!1\ge 1,\ \text{so by above}\\[.2em] &{\rm this\ \ is\,\ \underset{\textstyle\color{#c00}{unsolvable}}{ solvable}\:\ for}\ \ \,n\!+\!17\!+\!23\underset{\textstyle\color{#c00}{\bf =^{\phantom{-}\!\!\!\!}}}> 17(23)\ \ {\rm i.e.}\ \ n \underset{\textstyle\color{#c00}{\bf =^{\phantom{-}\!\!\!\!}}}> 17(23)\!-\!17\!-\!23 \end{align}$


Alternatively we can use this inductive proof. Let $\,f(x,y)=17x+23y.\,$ The Bezout identity is $\,f(-4,3)=1.\,$ Hence if $\,f(x,y) = \color{#0a0}n\,$ then $\,f(\color{#c00}{x-4},y+3) = f(x,y)+f(-4,3) = \color{#0a0}{n+1},\,$ and if $\,\color{#c00}{x-4\le 0}\,$ then we can make it $> 0$ by further $\rm\color{#c00}{adding\ (23,-17)},\,$ since $\,f(23,-17)=0,\,$ e.g.

$391 = f(\ \ \ 0,17),\ $ so adding $1 = f(-4,3)\,$ yields
$392 = f(\color{#c00}{-4},20) = f(19,3)$ by $\rm\color{#c00}{adding\ (23,-17)}$
$393 = f(\ 15,6)$
$394 = f(\ 11,9)$
$395 = f(\ \ \ 7,12)$
$396 = f(\ \ \ 3,15)$
$397 = f(\color{#c00}{-1},18) = f(22,1)$ by $\rm\color{#c00}{adding\ (23,-17)}$
$\qquad \vdots$

Bill Dubuque
  • 272,048