Read · 1 min

We want to solve the equation ax4+bx3+cx2+dx+e=0a x^4+b x^3+c x^2+d x+e=0, assuming that a≠0a\ne0 so that it really is a quartic equation. First, as with equations of lower degree, we divide through by the leading coefficient: x4+bax3+cax2+dax+ea=0x^4+\frac{b}{a} x^3+\frac{c}{a} x^2+\frac{d}{a} x+\frac{e}{a}=0

Now that the equation is monic, we can factor our quartic: If α\alpha, β\beta, γ\gamma, and δ\delta are the roots of our quartic, since it is monic we have: x4+bax3+cax2+dax+ea=(x−α)(x−β)(x−γ)(x−δ)x^4+\frac{b}{a} x^3+\frac{c}{a} x^2+\frac{d}{a} x+\frac{e}{a}=\left(x-\alpha\right)\left(x-\beta\right)\left(x-\gamma\right)\left(x-\delta\right) This holds true for all xx, so we can equate coefficients of powers of xx:

−α−β−γ−δ=ba-\alpha-\beta-\gamma-\delta=\frac{b}{a} so α+β+γ+δ=−ba\alpha+\beta+\gamma+\delta=-\frac{b}{a}

αβ+αγ+αδ+βγ+βδ+γδ=ca\alpha\beta+\alpha\gamma+\alpha\delta+\beta\gamma+\beta\delta+\gamma\delta=\frac{c}{a}

−αβγ−αβδ−αγδ−βγδ=da-\alpha\beta\gamma-\alpha\beta\delta-\alpha\gamma\delta-\beta\gamma\delta=\frac{d}{a} so αβγ+αβδ+αγδ+βγδ=−da\alpha\beta\gamma+\alpha\beta\delta+\alpha\gamma\delta+\beta\gamma\delta=-\frac{d}{a}

αβγδ=ea\alpha\beta\gamma\delta=\frac{e}{a}

Now comes the key to this derivation, which is making a convenient substitution. Specifically, since

\begin{pmatrix}

1 & 1 & 1 & 1 \ 1 & 1 & - 1 & -1 \ 1 & -1 & 1 & -1 \ 1 & -1 & -1 & 1

\end{pmatrix}

is an invertible matrix, we can substitute:

α=ϵ+ζ+η+θ\alpha=\epsilon+\zeta+\eta+\theta

β=ϵ+ζ−η−θ\beta=\epsilon+\zeta-\eta-\theta

γ=ϵ−ζ+η−θ\gamma=\epsilon-\zeta+\eta-\theta

δ=ϵ−ζ−η+θ\delta=\epsilon-\zeta-\eta+\theta

Note that the product of the signs on ζ\zeta, η\eta, and θ\theta is always 1; this will be important at the end.

This substitution allows us to solve for ϵ\epsilon immediately: α+β+γ+δ=4ϵ=−ba\alpha+\beta+\gamma+\delta=4\epsilon=-\frac{b}{a}, so ϵ=−b4a\epsilon=-\frac{b}{4a}

αβ+αγ+αδ+βγ+βδ+γδ=6ϵ2−2ζ2−2η2−2θ2=ca\alpha\beta+\alpha\gamma+\alpha\delta+\beta\gamma+\beta\delta+\gamma\delta=6\epsilon^2-2\zeta^2-2\eta^2-2\theta^2=\frac{c}{a} so ζ2+η2+θ2=3b2−8ac16a2\zeta^2+\eta^2+\theta^2=\frac{3b^2-8ac}{16a^2}

αβγ+αβδ+αγδ+βγδ=4(ϵ(ϵ2−ζ2−η2−θ2)+2ζηθ)=−da\alpha\beta\gamma+\alpha\beta\delta+\alpha\gamma\delta+\beta\gamma\delta=4\left(\epsilon\left(\epsilon^2-\zeta^2-\eta^2-\theta^2\right)+2\zeta\eta\theta\right)=-\frac{d}{a} so, since ϵ=−b4a\epsilon=-\frac{b}{4a} and ζ2+η2+θ2=3b2−8ac16a2\zeta^2+\eta^2+\theta^2=\frac{3b^2-8ac}{16a^2}, we obtain ζηθ=−8a2d+4abc−b364a3\zeta\eta\theta=\frac{-8a^2d+4abc-b^3}{64a^3}

ea=αβγδ=ϵ4+ζ4+η4+θ4−2ϵ2ζ2−2ϵ2η2−2ϵ2θ2−2ζ2η2−2ζ2θ2−2η2θ2+8ϵζηθ\frac{e}{a}=\alpha\beta\gamma\delta=\epsilon^4+\zeta^4+\eta^4+\theta^4-2\epsilon^2\zeta^2-2\epsilon^2\eta^2-2\epsilon^2\theta^2-2\zeta^2\eta^2-2\zeta^2\theta^2-2\eta^2\theta^2+8\epsilon\zeta\eta\theta so, since ϵ=−b4a\epsilon=-\frac{b}{4a} and ζηθ=−8a2d+4abc−b364a3\zeta\eta\theta=\frac{-8a^2d+4abc-b^3}{64a^3}, we obtain ϵ4+ζ4+η4+θ4−2ϵ2ζ2−2ϵ2η2−2ϵ2θ2−2ζ2η2−2ζ2θ2−2η2θ2=32a3e+b(−8a2d+4abc−b3)32a4\epsilon^4+\zeta^4+\eta^4+\theta^4-2\epsilon^2\zeta^2-2\epsilon^2\eta^2-2\epsilon^2\theta^2-2\zeta^2\eta^2-2\zeta^2\theta^2-2\eta^2\theta^2=\frac{32a^3e+b\left(-8a^2d+4abc-b^3\right)}{32a^4} Since ζ2+η2+θ2=3b2−8ac16a2\zeta^2+\eta^2+\theta^2=\frac{3b^2-8ac}{16a^2}, we can add ϵ2=b216a2\epsilon^2=\frac{b^2}{16a^2} to get ϵ2+ζ2+η2+θ2=b2−2ac4a2\epsilon^2+\zeta^2+\eta^2+\theta^2=\frac{b^2-2ac}{4a^2}.

Then squaring both sides gives ϵ4+ζ4+η4+θ4+2ϵ2ζ2+2ϵ2η2+2ϵ2θ2+2ζ2η2+2ζ2θ2+2η2θ2=(b2−2ac)216a4\epsilon^4+\zeta^4+\eta^4+\theta^4+2\epsilon^2\zeta^2+2\epsilon^2\eta^2+2\epsilon^2\theta^2+2\zeta^2\eta^2+2\zeta^2\theta^2+2\eta^2\theta^2=\frac{\left(b^2-2ac\right)^2}{16a^4}

Then subtracting gives us 4ϵ2ζ2+4ϵ2η2+4ϵ2θ2+4ζ2η2+4ζ2θ2+4η2θ2=(ϵ2+ζ2+η2+θ2)2−(ϵ4+ζ4+η4+θ4−2ϵ2ζ2−2ϵ2η2−2ϵ2θ2−2ζ2η2−2ζ2θ2−2η2θ2)=(b2−2ac)216a4−ea−b(−8a2d+4abc−b3)32a4=2(b2−2ac)2−32a3e+b(8a2d−4abc+b3)32a44\epsilon^2\zeta^2+4\epsilon^2\eta^2+4\epsilon^2\theta^2+4\zeta^2\eta^2+4\zeta^2\theta^2+4\eta^2\theta^2=\left(\epsilon^2+\zeta^2+\eta^2+\theta^2\right)^2-\left(\epsilon^4+\zeta^4+\eta^4+\theta^4-2\epsilon^2\zeta^2-2\epsilon^2\eta^2-2\epsilon^2\theta^2-2\zeta^2\eta^2-2\zeta^2\theta^2-2\eta^2\theta^2\right)=\frac{\left(b^2-2ac\right)^2}{16a^4}-\frac{e}{a}-\frac{b\left(-8a^2d+4abc-b^3\right)}{32a^4} =\frac{2\left(b^2-2ac\right)^2-32a^3e+b\left(8a^2d-4abc+b^3\right)}{32a^4} so we can now obtain ϵ2ζ2+ϵ2η2+ϵ2θ2+ζ2η2+ζ2θ2+η2θ2=2(b2−2ac)2−32a3e+b(8a2d−4abc+b3)128a4\epsilon^2\zeta^2+\epsilon^2\eta^2+\epsilon^2\theta^2+\zeta^2\eta^2+\zeta^2\theta^2+\eta^2\theta^2= \frac{2\left(b^2-2ac\right)^2-32a^3e+b\left(8a^2d-4abc+b^3\right)}{128a^4} Since ϵ=−b4a\epsilon=-\frac{b}{4a} and ζ2+η2+θ2=3b2−8ac16a2\zeta^2+\eta^2+\theta^2=\frac{3b^2-8ac}{16a^2}, ϵ2ζ2+ϵ2η2+ϵ2θ2=ϵ2(ζ2+η2+θ2)=b216a2⋅(3b2−8ac)16a2=b2(3b2−8ac)256a4\epsilon^2\zeta^2+\epsilon^2\eta^2+\epsilon^2\theta^2=\epsilon^2\left(\zeta^2+\eta^2+\theta^2\right)=\frac{b^2}{16a^2}\cdot\frac{\left(3b^2-8ac\right)}{16a^2}=\frac{b^2\left(3b^2-8ac\right)}{256a^4}

Then we can subtract to obtain ζ2η2+ζ2θ2+η2θ2=4(b2−2ac)2−64a3e+2b(8a2d−4abc+b3)+b2(8ac−3b2)256a4\zeta^2\eta^2+\zeta^2\theta^2+\eta^2\theta^2=\frac{4\left(b^2-2ac\right)^2-64a^3e+2b\left(8a^2d-4abc+b^3\right)+b^2\left(8ac-3b^2\right)}{256a^4}

This means ζ2\zeta^2, η2\eta^2, and θ2\theta^2 are the roots of the cubic (y−ζ2)(y−η2)(y−θ2)=y3−(ζ2+η2+θ2)y2+(ζ2η2+ζ2θ2+η2θ2)y−(ζηθ)2=y3+(8ac−3b2)16a2y2+4(b2−2ac)2−64a3e+2b(8a2d−4abc+b3)+b2(8ac−3b2)256a4y−(8a2d−4abc+b3)24096a6\left(y-\zeta^2\right)\left(y-\eta^2\right)\left(y-\theta^2\right)=y^3-\left(\zeta^2+\eta^2+\theta^2\right)y^2+\left(\zeta^2\eta^2+\zeta^2\theta^2+\eta^2\theta^2\right)y-\left(\zeta\eta\theta\right)^2=y^3+\frac{\left(8ac-3b^2\right)}{16a^2}y^2+\frac{4\left(b^2-2ac\right)^2-64a^3e+2b\left(8a^2d-4abc+b^3\right)+b^2\left(8ac-3b^2\right)}{256a^4}y-\frac{\left(8a^2d-4abc+b^3\right)^2}{4096a^6}. The importance of the product of signs in the substitutions defining ϵ\epsilon, ζ\zeta, η\eta, and θ\theta is that just four sign combinations give roots of our original quartic, and they are the ones making (±ζ)(±η)(±θ)=ζηθ=−8a2d+4abc−b364a3(\pm\zeta)( \pm\eta)( \pm\theta) = \zeta\eta\theta=\frac{-8a^2d+4abc-b^3}{64a^3}. So the solutions to our original quartic are given by x=−b±y1±y2±y34ax=\frac{-b \pm\sqrt{y_1}\pm\sqrt{y_2}\pm\sqrt{y_3}}{4a}, where the signs are chosen so that (±y1)(±y2)(±y3)=−8a2d+4abc−b3\left(\pm\sqrt{y_1}\right)\left(\pm\sqrt{y_2}\right)\left(\pm\sqrt{y_3}\right)=-8a^2d+4abc-b^3

(There's a reason this formula is almost never written explicitly and directly in terms of the coefficients of the original quartic.)

Type a concept, like “slope” or “photosynthesis”.

↑ ↓ to choose · Enter to openSearch explanations and problems