Understanding Cardano's method of solving Cubic equation

  • Context: Graduate 
  • Thread starter Thread starter PcumP_Ravenclaw
  • Start date Start date
  • Tags Tags
    Cubic Method
Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
3 replies · 3K views
PcumP_Ravenclaw
Messages
105
Reaction score
4
Dear All,
I am trying to understanding Cardano's method to solve the cubic equation. Please first see the reference text that is attached (Alan F. Beardon, Algebra and Geometry).

My rough understanding is that we make 2 substitutions (## P1(z - a/3)## and ##P(z - b/z) ## ) to simplify the cubic equation to the from (## z^6 -cz^3 - b^3 = 0 ##) so that we can make a further substitution (## y = z^3##). This results in a quadratic equation which can be solved to get two solutions y1 and y2.

## z^3 = y1## or ##y2 ## say y1 and y2 are 8 and -8 then this results in only two solutions right 2 and -2. Without forgetting that we need to do 2 Reverse substitutions to get the final solution. Why does the book say 6 solutions for z and only 3 are distinct?

In the paragraph below from the text.
" However, the values of ζ are the
roots of the equation ## z^6 − cz^3 − b^3 = 0## , and if v is a root of this equation,
then so (trivially) is v1 = −b/v. As v1 − b/v1 = v − b/v, we see that these six
values of ζ can only provide at most three distinct roots of p. "

what is v and v1. why is v1 = -b/v (The substitution was ##P(z - b/z) ## ) ??

I think the coefficients, a, b & c from the equation 3.6.1 and the equation below 3.6.1 in the text are NOT the same as the original cubic equation (## z^3 + az^2 + bz + c = 0 ##). I cross-referenced equation 3.6.1 with equation 12 from http://mathworld.wolfram.com/CubicFormula.html. I don't think the coefficients are the same. Are they??

Danke...
 

Attachments

  • 58979-f8da60824d2a9e7294210a5a6955bf60.jpg
    58979-f8da60824d2a9e7294210a5a6955bf60.jpg
    54.2 KB · Views: 836
Last edited by a moderator:
Physics news on Phys.org
PcumP_Ravenclaw said:
z3=y1ory2 z^3 = y1 or y2 say y1 and y2 are 8 and -8 then this results in only two solutions right 2 and -2. Without forgetting that we need to do 2 Reverse substitutions to get the final solution. Why does the book say 6 solutions for z and only 3 are distinct?
In the complex numbers, z^3 = y always has three distinct solutions (let's ignore the case y=0). Therefore, you get up to 6 solutions. At most 3 are relevant for the final solutions.

PcumP_Ravenclaw said:
what is v and v1.
Solutions to the equations discussed above.
PcumP_Ravenclaw said:
why is v1 = -b/v
Definition. It is like a statement "if n is an integer, then m=n+1 is an integer": "if v is a solution, then v1=-b/v is a solution" (check this!).

b and c are not the original b and c, right.
 
  • Like
Likes   Reactions: PcumP_Ravenclaw
Here is my understanding of Cardano's method: If a and b are any numbers then [itex](a+ b)^3= a^3+ 3a^2b+ 3ab^2+ b^3[/itex] and [itex]3ab(a+ b)= 3a^2b+ 3ab^2[/itex]. Subtracting, [itex](a+ b)^3- 3ab(a+ b)= a^3+ b^3[/itex].

Letting x= a+ b, m= 3ab, and [itex]n= a^3+ b^3[/itex] then x satisfies the reduced cubic [itex]x^3+ mx= n[/itex] ("reduced" because it has no squared term). Now, what about the other way around? That is, if we know m and n, can we find a and b and so x?

Yes, we can. From [itex]m= 3ab[/itex], we have [itex]b= \frac{m}{3a}[/itex] so that
[tex]n= a^3+ b^3= a^3+ \frac{m^3}{3^3a^3}[/tex]
Multiply by [itex]a^3[/itex]: [itex]na^3= (a^3)^2+ \left(\frac{m}{3}\right)^3[/itex] or [itex](a^3)^2- na^3+ \left(\frac{m}{3}\right)^3= 0[/itex].

We can think of that as a quadratic equation in [itex]a^3[/itex] and, by the quadratic formula
[tex]a^3= \frac{n\pm\sqrt{n^2- 4\left(\frac{m}{3}\right)^3}}{2}= \frac{n}{2}\pm\sqrt{\left(\frac{n}{2}\right)^2- \left(\frac{m}{3}\right)^2}[/tex]
From [tex]a^3+ b^3= n[/tex]
[tex]b^3= n- a^3= n- \frac{n}{2}\pm\sqrt{\left(\frac{n}{2}\right)^2- \left(\frac{m}{3}\right)^2}=\frac{n}{2}\mp\sqrt{\left(\frac{n}{2}\right)^2- \left(\frac{m}{3}\right)^2}[/tex]

Take cube roots to find a and b and then x= a+ b. There are, of course, three cube roots for each so 9 possible combinations but some add to give the same x.
 
  • Like
Likes   Reactions: PcumP_Ravenclaw and mfb