# How do mathematica and maple solve cubics?

• Mathematica
wumple
Hi,

Does anyone know how mathematica and maple solve cubics? Do they do it numerically?

Simon_Tyler
dreiter

Not to be rude, but that link by itself isn't terrifically helpful for him.

I don't know the answer to your question, but Mathematica can solve an equation numerically or algebraically, depending on the instructions you give it. Do you have a spefic problem in mind?

implicitnone
I once tried to learn the basic of symbolic computation but it's not easy, I quitted :) They use very advanced algebra theorems for convergences and so on. The numeric codes should be as complicated as the symbolic ones as well. They're all highly optimized.

Simon_Tyler
Not to be rude, but that link by itself isn't terrifically helpful for him.

I agree -- but so many times on this forum, you put in a lot of effort into an answer and never get a reply.

The OP posted a one line question which showed very little hint that they'd done any reading on the subject - so I assumed they were just wowed by the large algebraic mess that solving cubics gives. The wikipedia article gives a good discussion of the history and method.

If the OP then wanted to be more specific about numerics or say symbolic factorization over various rings, then they could have come back and clarified their interest.

wumple
Well my understanding is that solving a cubic algebraically comes down to trisecting an angle, which is of course not terribly doable...for example, I put in coefficients that should give me a real root of -1 for the depressed cubic

x^3 + ax + b = 0

but if I make the substitution x = u + v and go through the algebra, I come up with something of the form

x = ((1+z)^(1/3) + (1-z)^(1/3))

So how do these programs get around this issue?