2

The question says to find the 4th degree polynomial with integral coefficients whose root is $\sqrt{3}$ + $\sqrt{2}$

This is how I solved

I assumed $ax^{4} + bx^{3} + cx^{2} + dx + e$ to be the polynomial where a,b,c,d,e are integers.

as $x = \sqrt{3}$ + $\sqrt{2}$ is a root of the polynomial, thus putting $x$ in it yields 0 as the result.

This gives

$2\sqrt{6}(10a+c) + \sqrt{3}(9b+d) + \sqrt{2}(11b+d) + 49a + 5c + e = 0 \;\;\;$ ...(1)

Now, as $a,b,c,d,e$ are integers thus

$10a + c = 0$

$9b + d = 0$

$11b + d = 0$ and

$49a + 5c + e = 0$

for the equation (1) to be satisfied

By solving these 4 relations, we get

$b = d = 0$

$a = e$ and

$c = -10e$

Therefore our polynomial becomes

$ex^{4} - (10e)x^{2}+ e$

But now I am stuck with finding the value of $e$.

The answer to the problem is

$x^{4} - (10)x^{2}+1 $

but I think e can be any integer as there are no other constraints given in the question. Is the answer wrong or did I do any mistakes in the procedure?

Haris
  • 3,409
Raghav
  • 31
  • 6
  • 1
    This looks correct. Maybe they assumed the polynomial's leading coefficient is $1$. That would make a unique solution. – user3257842 Feb 15 '24 at 07:48
  • 1
    Yes, $e$ is arbitrary. You can see this by noting that your final polynomial is of the form $e\cdot P(x)$ for $P(x)=x^4-10x^2+1$. Then $P(x)$ and $e\cdot P(x)$ have exactly the same roots, regardless of the choice of $e$. – freakish Feb 15 '24 at 07:49
  • The usual question is this one, which has been asked a lot. – Dietrich Burde Feb 15 '24 at 12:07

3 Answers3

6

Start with the fact that $x=\sqrt 2+\sqrt 3$ is a solution to the polynomial. Squaring both sides gives $x^2=2+2\sqrt 6+3=5+2\sqrt 6$. Thus $x^2-5=2\sqrt 6$. Squaring again, $x^4-10x^2+25=24$. Now observe that the equation $x^4-10x^2+1=0$ is satisfied by $x=\sqrt 2+\sqrt 3$.

Your solution is also correct. Notice that there are infinitely many solutions, since $\sqrt 2+\sqrt 3$ remains a solution after you multiply by any integer. Once you reach the solution $ex^4-10ex^2+e$, you are free to choose $e=1$ to get $x^4-10x^2+1$. You could just as easily pick $e=-17$. (This is partly why you often see problems ask for a monic polynomial. Then there is less ambiguity since the leading coefficient must be $1$.)

pancini
  • 19,216
  • 1
    @freakish true I guess. Edited a bit. – pancini Feb 15 '24 at 08:07
  • I see. So basically, the question hasn't provided sufficient information. – Raghav Feb 15 '24 at 08:09
  • @Raghav that's right, but it's arguably a minor point. Depending on the context (if you are worried about finding integral extensions of rings, for example, which wouldn't come up before a university-level abstract algebra course), it doesn't really matter what the leading coefficient is. The question would be totally fine if you just changed "the 4th degree polynomial" to "a 4th degree polynomial," for example. – pancini Feb 15 '24 at 08:14
  • The question states "Give me an equation where $\sqrt{2}+\sqrt{3}$ is a solution", and you give the equation where $\sqrt{2}+\sqrt{3}$ is the solution. – Dominique Feb 15 '24 at 08:19
  • 1
    @Dominique no, but even if it were the only solution, $\sqrt 2+\sqrt 3$ would still be called "a solution." – pancini Feb 15 '24 at 08:20
2

A different point of view (leading in this case to a longer calculation, I admit heartily) is to add to the given root $x_0$ the list of its conjugate roots which are necessarily roots of the same (minimal) polynomial :

$$\begin{cases}x_0&=&+\sqrt{2}+\sqrt{3}\\ x_1&=&+\sqrt{2}-\sqrt{3}\\ x_2&=&-\sqrt{2}+\sqrt{3}\\ x_3&=&-\sqrt{2}-\sqrt{3}\\\end{cases}$$

and expand monic polynomial :

$$(x-x_0)(x-x_1)(x-x_2)(x-x_3)$$

$$=((x+\sqrt{2})^2-(\sqrt{3})^2)((x-\sqrt{2})^2-(\sqrt{3})^2)$$

$$=(x^2+2\sqrt{2}x-1)(x^2-2\sqrt{2}x-1)$$

$$=((x^2-1)+2\sqrt{2}x)((x^2-1)-2\sqrt{2}x)$$

$$=(x^2-1)^2-(2\sqrt{2}x)^2$$

$$=x^4-2x^2+1-8x^2$$

$$=x^4-10x^2+1$$

Besides, have a look at the reverse side of the medal : the irreducibility of the obtained polynomial as treated in the interesting answers to this question :

Show that $x^4-10x^2+1$ is irreducible over $\mathbb{Q}$

Remark : this kind of question has already been asked a certain number of times like here.

Jean Marie
  • 81,803
1

Hint: The minimal polynomial of $(\sqrt2+\sqrt3)^2=5+2\sqrt6$ is $x^2-10x+1$.

Bob Dobbs
  • 10,988
  • Interesting remark ; in view of the list of roots I gave in my answer, it is like grouping by $2$ : $x_0$ with $x_3=-x_0$ ; and $x_1$ with $x_2=-x_1$. – Jean Marie Feb 15 '24 at 12:25