When does quartic have a real root?

When does a quartic have a real root?

This is a method for deciding if a polynomial has any real roots (or equivalently when it has no real roots). Start with a quartic polynomial.
Ax4+4Bx3+6Cx2+4Dx+E = 0

or QnmopXnXmXoXp = 0

The (det_4(Q)) discriminant is:
\begin{align} 
\Delta\ =\ &256 a^3 e^3 - 192 a^2 b d e^2 - 128 a^2 c^2 e^2 + 144 a^2 c d^2 e - 27 a^2 d^4 \\ 
&+ 144 a b^2 c e^2 - 6 a b^2 d^2 e - 80 a b c^2 d e + 18 a b c d^3 + 16 a c^4 e \\
&- 4 a c^3 d^2 - 27 b^4 e^2 + 18 b^3 c d e - 4 b^3 d^3 - 4 b^2 c^3 e + b^2 c^2 d^2
\end{align}
also
D=A^3.E-4.A^2.B.D-9.A^2.C^2+24.A.B^2.C-12.B^4
D' = (D + P^2) = d^2/dE^2  det_4(Q)
also
P = AC-B^2  = discriminant(  d^2/dx^2 p(x))  

It has NO real roots when
det_4(Q) >0  and D>0 and P>0
or det_4(Q) =0 and D=0 and P>0


Question
How to get these polynomials as determinants of matrices? Or just find a simple algorithm. Guess: Are these the resultants of f(x) with it's derivatives? f'''''(x). How are D and P generalized?

Extensions
How to extend this to multi-variable polynomials? (This would solve kissing number problem). Obviously, we could have the coefficients depend on values of y; a(y), b(y), c(y), etc. Then we would have a series of inequalities in y. But is there a better way?
If we can solve any set of functions f(x)>0 , g(x)>0, h(x)>0, ...we can solve for any variables.

For quadratic case multi-variable polynomial is simply d^T.c^-1.d - e >0


Appendix
a>0 & b>0 <==> a+b>0 & ab>0 
a>0 &  b>0 & c>0   <==> a+b+c>0 & ab+bc+ca>0 & abc>0

Inequality

ax4+bx3+cx2+dx+e > 0 


has no solution if it has no roots and also a<0.


The equality is equivalent to:

x2- y = 0
ay2+bxy+cnx2+c(a-n)y + dx+e = 0

General Question
When can an equation f(x,y,z,....)=0 be solved in real variables?



Kissing number Problem
Let (x_n-x_m)^2 = R_nm^2 = 1 + (S_nm)^2
Area squared of simplex in n+1 dimensions =  A(1,2,3,...,n,k)=f(S)

Equation to solve in real variables for S:
Sum A(1,2,3,..,n,k)^k = 0

In the 1D case p=S12, q=S23, r=S13

((1+p2)2+(1+q2)2+(1+r2)2 - 2(1+p2)(1+q2)- 2(1+p2)(1+r2)- 2(1+r2)(1+q2))2 = 0 has no solution

1D: 3 variables, 8th degree
2D: 21 variables, 12th degree?
3D: 78 variables, 16th degree?


In general f(x)>0 an nth power polynomial requires (n-1) conditions.


ax4+bx3+cx2+dx+e = 0

=> (y-x2)2+(ay2+byx+cx2+dx+e)2  =  0
y^4 (a) + y^3x ( 2ab)+ y^2x^2(2ac) + yx^3(2bc)+x^4(1+c^2) + ..... = 0

Which is a quartic!! But in 2 variables.


Conclusion


Solving multi-variable quartics solves ALL systems of polynomial equations.

Proof: A system of quadratics can be written in the form:(XTAX)2+(XTBX)2+(XTCX)2+(XTDX)2+... = 0

AnmiApqi + (XAX)(bX) AnmiBi


All systems of polynomial equations and non-strict inequalities can easily be reduced to the following form:

AnmopXXXX+BnmoXXX+CnmXX+DnX+E = 0

or more precisely:

AnmiAopiXXXX = 0
where X=(x,y,z,...1)

e.g. Ex{a+bx>=0}  => Exy{y^2-ax-b = 0}

what is integral exp(AX^4+BX^3+CX^2+D)dX?


Det(B)=

[00]+[01]
[11]+[10]

[00]+[00]+[10]+[20]
[11]+[12]+[21]+[01]
[22]+[21]+[02]+[12]



Comments

Popular posts from this blog

Vortex Solution to Navier Stokes Equations

Conjugate of Lie Algebra