Solving a cubic equation: 2x3 - 3x - 5

TheOAP

New member
Joined
Nov 5, 2017
Messages
16
In a program I wrote the equation came up with "No solution" :
2x3 - 3x - 5 using Synthetic Division. Tried 'manually' - same answer!
Went to a (now defunct) forum and was shown a technique :
Assume the equation is ax3 + bx2 +cx +d
Let x = y - b/3a
which is then shown expanded.

Then says multiplying out and simplifying" to get :

a.y^3 + ( c - b^2 / 3a )y + ( d + 2.b^3 / 27.a^2 - b.c/3a) = 0

I've tried but cannot get this!

Hope you can help as I would like this technique to add to my program.
Pete
 
In a program I wrote the equation came up with "No solution" :
2x3 - 3x - 5 using Synthetic Division. Tried 'manually' - same answer!
Went to a (now defunct) forum and was shown a technique :
Assume the equation is ax3 + bx2 +cx +d
Let x = y - b/3a
which is then shown expanded.

Then says multiplying out and simplifying" to get :

a.y^3 + ( c - b^2 / 3a )y + ( d + 2.b^3 / 27.a^2 - b.c/3a) = 0

I've tried but cannot get this!

Hope you can help as I would like this technique to add to my program.
Pete
What you are remembering, but not quite accurately, is Cardano's method for solving the GENERAL cubic. The substitution that you are remembering is used to eliminate the squared term in the general cubic to get the depressed cubic. But your function does not have a squared term; it already is a DEPRESSED cubic so that substitution is irrelevant. Then, if I remember correctly, there is another set of substitutions applied to the depressed cubic which gives a simplified sixth degree polynomial that can be solved using the quadratic formula.

The only purpose for Cardano's method today is to prove the cubic formula. So why not just use the cubic formula in your program?

Here is a site on Cardano's method, but it may hint that a better method is suitable for execution on a computer.

https://www.math.ucdavis.edu/~kkreith/tutorials/sample.lesson/cardano.html
 
Last edited:
Cubic Polynomial. There IS at least ONE Real Solution. Better rewrite your program.
 
Solving cubic equations

So equation is 2x^3 - 3x - 5 = 0
There is a real solution, plus 2 complex solutions.
Real solution is ~1.7189

So something's wrong with your program...

Hi Denis,
I meant to "unravel" the equation after depression!

a( y - b/3a)^3 + b( y - b/3a)^2 + c( y - b/3a) + d

It says multiply out and simplify to get :

ay^3 + (c - (b^2/3a))*y + ( d + (2b^3/27a^2) - bc/3a)
My excuse is that I have poor vision and I'm 82 and slightly GaGa!

Regards
 
Solving cubic equations

What you are remembering, but not quite accurately, is Cardano's method for solving the GENERAL cubic. The substitution that you are remembering is used to eliminate the squared term in the general cubic to get the depressed cubic. But your function does not have a squared term; it already is a DEPRESSED cubic so that substitution is irrelevant. Then, if I remember correctly, there is another set of substitutions applied to the depressed cubic which gives a simplified sixth degree polynomial that can be solved using the quadratic formula.

The only purpose for Cardano's method today is to prove the cubic formula. So why not just use the cubic formula in your program?

Here is a site on Cardano's method, but it may hint that a better method is suitable for execution on a computer.

https://www.math.ucdavis.edu/~kkreith/tutorials/sample.lesson/cardano.html

Hi JeffM

a.y^3 + ( c - b^2 / 3a )y + ( d + 2.b^3 / 27.a^2 - b.c/3a) = 0
It says multiply out and simplify to get :
ay^3 + ( c - (b^2 / 3a))*y + ( d + ( 2b^3 / 27a^2 ) - bc /3a) = 0
but I cant "unravel" it!
My excuse : I'm 82 and slightly GaGa!
Pete
 
Hi JeffM

a.y^3 + ( c - b^2 / 3a )y + ( d + 2.b^3 / 27.a^2 - b.c/3a) = 0
It says multiply out and simplify to get :
ay^3 + ( c - (b^2 / 3a))*y + ( d + ( 2b^3 / 27a^2 ) - bc /3a) = 0
but I cant "unravel" it!
My excuse : I'm 82 and slightly GaGa!
Pete
I am 73 and totally gaga. You say "it" says but do not quote what it says. So I am not sure whether "it" is wrong or whether you are misreading it. As far as I can remember the substitution you are that trying to apply is not relevant to a depressed cubic, which is what you seem to be working on.
\(\displaystyle a \ne 0 \text { and } ax^3 + bx^2 + cx + d = 0 \implies x^3 + x^2 * \dfrac{b}{a} + x * \dfrac{c}{a} + \dfrac{d}{a} = 0.\)

\(\displaystyle \text {Let } y = x + \dfrac{b}{3a} \implies x = y - \dfrac{b}{3a}.\)

\(\displaystyle \therefore x^3 = y^3 - 3y^2 * \dfrac{b}{3a} + 3y * \dfrac{b^2}{9a^2} - \dfrac{b^3}{27a^3} = y^3 - \dfrac{by^2}{a} + \dfrac{b^2y}{3a^2} - \dfrac{b^3}{27a^3}.\)

\(\displaystyle \text { And } x^2 * \dfrac{b}{a} = \dfrac{b}{a} \left (y^2 - 2y * \dfrac{b}{3a} + \dfrac{b^2}{9a^2} \right ) = \dfrac{by^2}{a} - \dfrac{2yb^2}{3a^2} + \dfrac{b^3}{9a^3}.\)

\(\displaystyle \text {And } x * \dfrac{c}{a} = \dfrac{c}{a} \left ( y - \dfrac{b}{3a} \right ) = \dfrac{cy}{a} - \dfrac{bc}{3a^2}.\)

\(\displaystyle \therefore x^3 + x^2 * \dfrac{b}{a} + x * \dfrac{c}{a} + \dfrac{d}{a} = \)

\(\displaystyle y^3 - \dfrac{by^2}{a} + \dfrac{by^2}{a} + y * \left ( \dfrac{b^2}{3a^2} - \dfrac{2b^2}{3a^2} + \dfrac{c}{a} \right ) - \dfrac{b^3}{27a^3} + \dfrac{b^3}{9a^3} - \dfrac{bc}{3a^2} + \dfrac{d}{a} = \)

\(\displaystyle y^3 + y * \dfrac{3ac- b^2}{3a^2} + \dfrac{2b^3 - 9abc + 27a^2d}{27a^3} = 0.\).

\(\displaystyle \text {Let } b' = \dfrac{3ac - b^2}{3a^2} \text { and } c' = \dfrac{2b^3 - 9abc + 27a^2d}{27a^3} \implies y^3 + b'y + c' = 0.\)

In other words, you can always turn a general cubic into a depressed cubic missing a term in the unknown squared.
 
Last edited:
Cubic Polynomial. There IS at least ONE Real Solution. Better rewrite your program.

Hi tkhunny,
One Real solution - yes, but when using Synthetic Division only rational roots are produced.
JeffM gave me the info I needed
Thanks
 
Solving a cubic equatiion: 2x3 - 3x - 5

Hi Denis,
I meant to "unravel" the equation after depression!

a( y - b/3a)^3 + b( y - b/3a)^2 + c( y - b/3a) + d

It says multiply out and simplify to get :

ay^3 + (c - (b^2/3a))*y + ( d + (2b^3/27a^2) - bc/3a)
My excuse is that I have poor vision and I'm 82 and slightly GaGa!

Regards
Hi Denis,
Try Synthetic Div. yourself. It wont produce a rational root!
I'm going to try "inching" my way to the nearest solution - it should be interesting..
So far:
Forcing a factor of 1.71885943 => remainder of 0.000085558591689682828
Pete
 
Cubic - irrational root

So equation is 2x^3 - 3x - 5 = 0
There is a real solution, plus 2 complex solutions.
Real solution is ~1.7189

So something's wrong with your program...

Have extended my program to catch an irrational root.
As the result is irrational then the other roots cant be found - surely?
(x + Something...)(x - Something else...).
Something and Something else aren't calculable so
I display the approx root plus the range of the remainder
approaching from the +ive & -ive sides.
This is what is displayed :

No rational solution found
Root is approximately
1.71885000000315
Remainder after Synthetic Division is between :
-5.33147953252211E-05 and 9.39529553374285E-05

Any useful comments ?
Regards
 
Have extended my program to catch an irrational root.
As the result is irrational then the other roots cant be found - surely?
(x + Something...)(x - Something else...).
Something and Something else aren't calculable so
I display the approx root plus the range of the remainder
approaching from the +ive & -ive sides.
This is what is displayed :

No rational solution found
Root is approximately
1.71885000000315
Remainder after Synthetic Division is between :
-5.33147953252211E-05 and 9.39529553374285E-05

Any useful comments ?
Regards

Synthetic Division absolutely does NOT capture only Rational Roots. It will find ALL of them, no matter how many or what type.

You do know that Synthetic Division is just evaluating the polynomial, right? It's called "Horner's Method". Look it up.

Actually, the operational manipulation provides more information. If you track the intermediate numbers, you will not need to go find them when you decide to tackle the rest of the roots.
 
Hi you can solve it by Cardano formula
ax3+bx2+cx+d=0
Cardano Formula
y3+py+q=0,x=y−b3a, p=−b23a2−ca, q=2b327a3−bc3a2+dax=−q2+−Δ108−−−−−√−−−−−−−−−−−−√3+−q2−−Δ108−−−−−√−−−−−−−−−−−−√3, Δ=−27q2−4p3
a≠0⟹x1=1.72
you can check this calculator for solving cubic equations
 
Top