# I 5th order equation unsolvable by approximation methods?

1. Mar 15, 2016

### arupel

I am accross this equation with no known analytical solutions:

x^5 - x + 1 = 0.

I asked this before and the answer was that you can get as good an approximation you want by approximation methods.

It is possible that when applying approximation methods, you will get singularities. These methods won't work. The equation is bullet proof.

The question is can we be cetain that there are approximation methods that will work or is this equation intrincally unvolvable even by approximatiuon methods?

My second question is messy. It is a stray thought.
Is this equaton one that satisified Godel's criteria for an unsolveable equation is tue but can be shown to be not provable and not unproveable? If it fits Godel's criteria it would be nice to show studends.

PS: Anyway, In teaching algebra, it is nice to show this unsolveable simple equation.

2. Mar 15, 2016

### jbriggs444

Binary search always works [on continuous functions defined over a contiguous range, given a pair of starting points]. If you can find a point where the graph of the function is positive and a point where the graph is negative then a binary search will converge toward a point where it is zero. By inspection, this polynomial is negative at x=-1000 and positive at x=+1000. Like all polynomials it is continuous and is defined over a contiguous range.

No, this has nothing to do with Godel.

Edit: Added some caveats to dot the i's and cross the t's

3. Mar 15, 2016

### Ssnow

Starting with the interval $[-2,-1]$, the solution is inside and you can approximate it with bisection method (for example) ... it seems to work ...

4. Mar 15, 2016

### pwsnafu

If you want any example for Godel see Goodstein's[/PLAIN] [Broken] theorem. Note that Goodstein can be expressed in PA, but not provable in PA. It is provable in ZFC.

Last edited by a moderator: May 7, 2017
5. Mar 15, 2016

### mathman

Galois proved that polynomial equations (in general) higher than 4th degree cannot be solved analytically.

6. Mar 15, 2016

### micromass

Staff Emeritus
That's a really inaccurate statement of what has actually been proved.
First, it's Abel and Ruffini who proved it first, not Galois. Although Galois' proof is superior.
Second, they are unsolvable using radicals, not with other analytic techniques. There are in fact some analytic techniques which can solve all 5 degree equations.
Third, this thread is about approximation, for which Galois theory is completely irrelevant.

7. Mar 15, 2016

### Thecla

A similar equation is discussed by Julian Havil in his book ,The Irrationals- A Story of the Numbers You Can't Count On". The equation is
x^5- x- 1=0. Julian says "no finite expression formed of the composition of radicals will provide an explicit expression for its roots, with its only real root
equal to 1.1673039782614186843...". p.133
Also the innocent looking equation: x^5- 5x+12=0 requires about 600 symbols for its exact expression in radicals for the only one real solution -1.842085966... p.133

8. Mar 15, 2016

### arupel

Thanks. I was hoping for a mystery; a continuous function which is absolutely unsolveble either analyticaly or by approximation methods, a platonic equation.

Sadly the mundane fact is that if it is a continious function it can be plotted. If it can be plotted it can be solved by the graph or analytically or by approximation methods, an obvious fact.

signing off