Re: Finding a continuous solution of a cubic

• To: mathgroup at smc.vnet.net
• Subject: [mg86310] Re: Finding a continuous solution of a cubic
• From: "David W.Cantrell" <DWCantrell at sigmaxi.net>
• Date: Sat, 8 Mar 2008 05:40:51 -0500 (EST)
• References: <fq3agu\$fq0\$1@smc.vnet.net> <fqb92v\$n6s\$1@smc.vnet.net> <fqlmnf\$jcm\$1@smc.vnet.net>

```Hugh Goyder <h.g.d.goyder at cranfield.ac.uk> wrote:
> Thanks to Andrzei and Daniel for their helpful comments -they have
> enabled me to simplify the problem. The response from David appears
> to  do exactly what I want but I am completely mystified as to what he
> did to get these  region boundaries. Please could you reveal the magic
> you used to find the regions?

I'm glad that I was able to give you exactly what you wanted. Namely,

"...the three functions you want, the solutions of d == 0,
can be expressed as

x -> Root[-2 a b + (-1 + b + a^2 b) #1 + #1^3 &, 1]
applicable when  27 a^2 b^2 <= (1 - b - a^2 b)^3 Or a <= 0

x -> Root[-2 a b + (-1 + b + a^2 b) #1 + #1^3 &, 2]
applicable when  27 a^2 b^2 <= (1 - b - a^2 b)^3

x -> -Root[2 a b + (-1 + b + a^2 b) #1 + #1^3 &, 1]
applicable when  27 a^2 b^2 <= (1 - b - a^2 b)^3 Or a >= 0 "

As to the boundaries, I hate to say that I merely graphed the solutions,
thought about the discriminant, and then wrote down the appropriate
conditions for applicability. In a private email, you said that you had
other similar problems and therefore hoped that I would disclose my method.
I've done so now, but it's not really a "method" in the sense you wanted.
Perhaps one of our more algebraically inclined cognoscenti could describe a
true "method".

David

```

• Prev by Date: RE: Increase length of arrows in vectorfieldplot3d
• Next by Date: Re: Rotation of 3D objects
• Previous by thread: Re: Finding a continuous solution of a cubic
• Next by thread: Re: Re: Re: Version 6.0.2