See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/262935349

A universal method of solving quartic equations

Article in International Journal of Pure and Applied Mathematics · January 2011

CITATIONS READS

89 5,092

1 author:

Sergei L. Shmakov
Saratov State University
34 PUBLICATIONS 176 CITATIONS

SEE PROFILE

All content following this page was uploaded by Sergei L. Shmakov on 09 June 2014.

The user has requested enhancement of the downloaded file.

International Journal of Pure and Applied Mathematics
————————————————————————–
Volume 71 No. 2 2011, 251-259

A UNIVERSAL METHOD OF SOLVING QUARTIC EQUATIONS

Sergei L. Shmakov
Saratov State University
83, Astrakhanskaya Str.,
Saratov, 410012, RUSSIAN FEDERATION

Abstract: All the existing methods of solving quartic equations (Descartes-

Euler-Cardano’s, Ferrari-Lagrange’s, Neumark’s, Christianson-Brown’s, and Ya-
coub-Fraidenraich-Brown’s ones) are particular versions of some universal method.
It enables producing an arbitrary number of other particular algorithms with
some desired properties.

Dedicated to Vera P. Filinova

my mathematics teacher.

AMS Subject Classification: 11D25

Key Words: quartic equation, solution, method, resolvent, shift, hyperbolic
transformation

Introduction

Quartic equations (often referred to as just “quartics”) in one unknown

x4 + ax3 + bx2 + cx + d = 0 (1)

are the highest degree polynomials which can be solved analytically, by radicals,
with no iterative techniques. There are several algorithms for this purpose, each
involving a subsidiary cubic equation, the so-called “resolvent”. In his paper
[1] Herbison-Evans examines them in application to computer graphics tasks.
Let us list these five algorithms with their resolvents.

Received: June 2, 2011 c 2011 Academic Publications, Ltd.

252 S.L. Shmakov

1. Descartes-Euler-Cardano’s one [2]:

y 3 + (2b − 3a2 /4)y 2 + (3a4 /16 − a2 b + ac + b2 − 4d)y−

 3 2
a ab
− − + c = 0; (2)
8 2

2. Ferrari-Lagrange’s [3]:

y 3 + by 2 + (ac − 4d)y + (a2 d + c2 − 4bd) = 0; (3)

3. Neumark’s [4]:

y 3 − 2by + (ac + b2 − 4d)y + (a2 d − abc + c2 ) = 0; (4)

4. Christianson-Brown’s: in 1967 Brown [5] published an original method of

solving palindromic quartics, and in 1991 Christianson [6] extended it to
a general quartic equation:

4a2 b − 4b2 − 4ac + 16d − 3a4 /4 2

y3 + y +
a3 − 4ab + 8c
+ (3a2 /16 − b/2)y + ab/16 − a3 /64 − c/8) = 0; (5)

5. Yacoub-Fraidenraich-Brown’s: Yacoub and Fraidenraich [7] have shown a

different way of applying Brown’s technique to a general quartic:

(a3 − 4ab + 8c)y 3 + (a2 b − 4b2 + 2ac + 16d)y 2 +

+ (a2 c − 4bc + 8ad)y + (a2 d − c2 ) = 0. (6)

By Herbison-Evans’ analysis [1], none of these five would completely suit

the needs of stable computations. He writes: “It would be good to find more
algorithms or variations on the five listed here which allowed other combinations
of coefficient signs to be handled in a stable fashion. It is far from clear why
there are five and only five algorithms so far discovered. Are there more? Are
there an infinite number of algorithms? ”
Let us try to answer these questions.
The basic algorithm is, as we will see, Ferrari-Lagrange’s one (with a minor
change).
 A UNIVERSAL METHOD OF SOLVING QUARTIC EQUATIONS 253

1. Basic Algorithm

Represent Eq. (1) as the product of two quadratic trinomials

(x2 + g1 x + h1 )(x2 + g2 x + h2 ) =
= x4 + (g1 + g2 )x3 + (g1 g2 + h1 + h2 )x2 + (g1 h2 + g2 h1 )x + h1 h2 ,

i.e. 

 g1 + g2 = a,
g1 g2 + h1 + h2 = b,

(7)
g h + g2 h1 = c,
 1 2


h1 h2 = d.

These pair sums and products of g and h allow Vièta’s theorem [8] to be applied
to express these coefficients as the roots of the subsidiary quadratic equations

g2 − ag + b − y = 0,

(8)
h2 − yh + d = 0,

where y is derived from the only equation with the cross products g1 h2 and
g2 h1 . Having explicitly written the roots, substituted them to the equation for
c, and eliminated the radical, one obtains the cubic equation for y (Appendix,
lines 1–9):
y 3 − by 2 + (ac − 4d)y − a2 d − c2 + 4bd = 0. (9)

This is the cubic resolvent of Eq. (1). In our case, it is identical to that of
Ferrari-Lagrange’s method, except for the sign at the even powers of y.
The algorithm is as follows:

1. Calculate the coefficients of Eq. (9) and find its real root (any cubic has
at least one real root).

2. Calculate the coefficients of Eqs (8) and solve them to get g1 , g2 , h1 , and
h2 . Eq. (1) is therefore factored into two quadratics.

3. Solve these quadratics.

This algorithm allows variations to be considered below.

254 S.L. Shmakov

2. A Shift of y

Introduce a new variable ys ≡ y − δ (i.e. y = ys + δ), where δ is an arbitrary

shift. For dimension reasons, it must look as δ = αa2 + βb, where α and β
are dimensionless coefficients; more sophisticated terms like ak bl cm dn , where
k + 2l + 3m + 4n = 2, should not be involved. The resolvent takes the form:

(ys + δ)3 − b(ys + δ)2 + (ac − 4d)(ys + δ) − a2 d − c2 + 4bd =

= ys3 + Ays2 + Bys + C = 0, (10)

where 
 A = 3δ − b,
B = 3δ2 − 2bδ + ac − 4d, (11)
C = δ3 − bδ2 + (ac − 4d)δ − a2 d − c2 + 4bd


(Appendix, lines 10–11). On finding a real root ys , a backward shift is due with
return to the rest of the master algorithm.
In Neumark’s method δ = b, and

 A = 3b − b = 2b,
B = 3b2 − 2b2 + ac − 4d = ac + b2 − 4d, (12)
C = b3 − b3 + (ac − 4d)b − a2 d − c2 + 4bd = abc + c2 − a2 d


(Appendix, line 12), and in Ferrari-Lagrange’s one, δ = 0, but ys in both

cases is negated, so the resolvent’s coefficients at the even powers of ys reverse
their signs, along with the sign at ys in Eqs (8).

3. A Shift of x

Introduce a new variable xs ≡ x − ∆ (i.e. x = xs + ∆), where ∆ is an arbitrary

shift. For dimension reasons, it must look as ∆ = γa, where γ is a dimensionless
coefficient; more sophisticated terms like ak bl cm dn , where k + 2l + 3m + 4n = 1,
should not be involved. Derive the coefficients of the initial quartic with respect
to xs :

(xs + ∆)4 + a(xs + ∆)3 + b(xs + ∆)2 + c(xs + ∆) + d =

= x4s + (a + 4∆)x3s + (b + 3a∆ + 6∆2 )x2s + (c + 2b∆ + 3a∆2 + 4∆3 )xs +
+ d + c∆ + b∆2 + a∆3 + ∆4 = x4s + as x3s + bs x2s + cs xs + ds = 0,
A UNIVERSAL METHOD OF SOLVING QUARTIC EQUATIONS 255

where as = a + 4∆, bs = b + 3a∆ + 6∆2 , cs = c + 2b∆ + 3a∆2 + 4∆3 , ds =

d + c∆ + b∆2 + a∆3 + ∆4 (Appendix, line 16). Substituting these expressions
to Eqs (11) for simultaneous application of both shifts, we have:
A = 3δ − bs = −b − 3a∆ + 3δ − 6∆2 ,


B = 3δ2 − 2bs δ + as cs − 4ds =





= 3δ2 − 2(b + 3a∆ + 6∆2 )δ + (a + 4∆)(c + 2b∆ + 3a∆2 +





+ 4∆3 ) − 4(d + c∆ + b∆2 + a∆3 + ∆4 ),





C = δ3 − bs δ2 + (as cs − 4ds )δ − a2s ds − c2s + 4bs ds = (13)
3 2 2 2
= δ − (b + 3a∆ + 6∆ )δ + ((a + 4∆)(c + 2b∆ + 3a∆ +





+ 4∆3 ) − 4(d + c∆ + b∆2 + a∆3 + ∆4 ))δ − (a + 4∆)2 (d+





+ c∆ + b∆2 + a∆3 + ∆4 ) − (c + 2b∆ + 3a∆2 + 4∆3 )2 +





+ 4(b + 3a∆ + 6∆2 )(d + c∆ + b∆2 + a∆3 + ∆4 ).


Calculations show that when δ = αa2s + βbs , where 8α + 3β = 1, the

coefficients of the resolvent do not change upon any shift of x. In Descartes-
Euler-Cardano’s method, the shift is chosen just in this way: δ = bs − a2s /4
(α = −1/4, β = 1). Additionally, ∆ = −a/4, so δ = b − 3a2 /8 and
3a2


 A = 2b − ,


 4
4
 3a
B= − a2 b + ac + b2 − 4d, (14)
 16
 3 2
a ab



 C=−
 − +c
8 2
(Appendix, lines 18–20).

4. Hyperbolic Transformation of y

Let’s assign the subscripts CB and YFB to the roots of Christianson-Brown’s

and Yacoub-Fraidenraich-Brown’s resolvents, respectively, that of the DEC re-
solvent remaining unsubscripted. Then
a3 + 8c − 4ab a3 + 8c − 4ab
y= =
16 yY F B + a4


16yCB
(note that the numerator (a3 + 8c − 4ab) is invariant with respect to the shift
∆ of x [9]). Substitute these expressions into Descartes-Euler-Cardano’s resol-
vent (14), eliminate the denominator (with possible loss of roots), and come to
the known coefficients (Appendix, lines 21–25).
256 S.L. Shmakov

Instead of immediate recalculation of their yCB and yY F B into y with re-

ducing the task to that already solved, the authors of the palindromic methods
propose their own sets of formulae, which hides the relation between several
forms of representation of the same, in essence, resolvent.

5. A New Particular Method

Having the described general scheme, one can easily produce an arbitrary num-
ber of particular algorithms. For example, let us require that the cubic resolvent
should contain no y 2 term (A = 0), then it follows from Eqs (11) that δ = b/3
(note that the condition 8α+3β = 1 holds true and the coefficients of the future
resolvent will be invariant to ∆) provided ∆ = 0, and
 2
b2

b b
 B=3 − 2b + ac − 4d = ac − − 4d,


 33  32 3



b b b

C= −b + (ac − 4d) − a2 d − c2 + 4bd = (15)


 3 3 3
abc 2 8


= − a2 d − b3 − c2 + bd


3 27 3
(Appendix, line 13). The subsidiary quadratics will be
2

 g2 − ag + b − ys = 0,

 3  (16)
2 b
 h − ys +
 h + d = 0,
3
their roots are
p
a± a2 − 8b/3 + 4ys
g1,2 = , (17)
p2
ys + b/2 ± (ys + b/3)2 − 4d
h1,2 = , (18)
2
(19)

and, finally, the roots of the initial quartic are

p
−g1 ± g12 − 4h1
x1,2 = , (20)
p2
−g2 ± g22 − 4h2
x3,4 = . (21)
2
A UNIVERSAL METHOD OF SOLVING QUARTIC EQUATIONS 257

A “new” algorithm for solving quartics is thus “invented”, it provides the cubic
resolvent just in its reduced form.

6. Conclusion

Of course, nobody could guarantee computational success to any of an arbitrary

number of particular algorithms. We may guess that the five previously pub-
lished methods have been elaborated so as to attach some algebraic quality to
the resolvent’s coefficients, but their robustness in actual computations should
be a subject of further analysis.
Excessive complications may be adverse. E.g., if in the palindromic algo-
rithms the expression (a3 + 8c − 4ab) takes on zero or a near-zero value, this
will cause a fault or an error. The case of yCB or (yY F B + a/4) close to zero is
also unfavorable. Simple shifts of the roots seem more safe, provided that no
close values are subtracted.
A similar technique of shifts and hyperbolic transformations can be applied
to cubic equations as well.

Acknowledgments

The author is grateful to Professor Donald Herbison-Evans for his encourage-

ment and advice.

References

[1] D. Herbison-Evans, Solving quartics and cubics for graphics, 2005

http://linus.it.uts.edu.au/~don/pubs/solving.html

[2] T. Strong, Elementary and Higher Algebra, Pratt and Oakley (1859).

[3] H.W. Turnbull, Theory of Equations, fourth ed., Oliver and Boyd, London
(1947).

[4] S. Neumark, Solution of Cubic and Quartic Equations, Pergamon Press,

Oxford (1965).

[5] K.S. Brown, Reducing Quartics to Cubics, 1967,

http://mathpages.com/home/kmath296.htm
258 S.L. Shmakov

[6] B. Christianson, Solving Quartics using Palindromes, The Math. Gaz. 75

(1991), 327–328.

[7] M.D. Yacoub, G. Fraidenraich. A New Simple Solution of the General

Quartic Equation, The Math. Gaz. 2011 (in press).

[8] F. Viète. Opera mathematica (1579). Reprinted Leiden, Netherlands

(1646).

[9] Y. Mochimaru, Reciprocal Solution of a Quartic Equation, Int. J. Pure

Appl. Math. 14 (2004), 207–210.

Appendix: MAXIMA’s script

(%i1) g^2-a*g+b-y$ solve(%,g)$

(%i3) h^2-y*h+d$ solve(%,h)$
(%i5) rhs(%th(3)[1])*rhs(%th(1)[2])+rhs(%th(3)[2])*
rhs(%th(1)[1])-c,expand;
p p
4 y − 4 b + a2 y 2 − 4 d ay
(%o5) − + −c
2 2
(%i6) pickapart(%,1);
p p
4 y − 4 b + a2 y 2 − 4 d
(%t6) −
2
ay
(%t7)
2
(%o7) − c + %t7 + %t6
(%i8) expandwrt(expand(%t6^2-(c-%t7)^2),y);
(%o8) y 3 − b y 2 − 4 d y + a c y + 4 b d − a2 d − c2
(%i9) create_list(expand(coeff(%,y,n)),n,[3,2,1,0]);
(%o9) [1, −b, a c − 4 d, 4 b d − a2 d − c2 ]
(%i10) subst([y=ys+D],%th(2)),expand;
(%o10) D3 + 3 ys D2 − b D2 + 3 ys2 D − 2 b ys D − 4 d D + a c D + ys3 − b ys2 − 4 d ys + a c ys +
4 b d − a2 d − c2
(%i11) [A,B,C]:create_list(coeff(%,ys,n),n,[2,1,0]);
(%o11) [3 D − b, 3 D2 − 2 b D − 4 d + a c, D3 − b D2 − 4 d D + a c D + 4 b d − a2 d − c2 ]
(%i12) subst([D=b],%),eval,expand;
(%o12) [2 b, −4 d + a c + b2 , −a2 d − c2 + a b c]
(%i13) subst([D=b/3],%th(2)),eval,expand;
b2 8 b d abc 2 b3
(%o13) [0, −4 d + a c − , − a2 d − c2 + − ]
3 3 3 27
(%i14) x^4+a*x^3+b*x^2+c*x+d$ subst([x=xs+D],%),expand$
(%i16) [as,bs,cs,ds]:create_list(coeff(%,xs,n),n,[3,2,1,0]);
(%o16) [4 D + a, 6 D2 + 3 a D + b, 4 D3 + 3 a D2 + 2 b D + c, D4 + a D3 + b D2 + c D + d]
(%i17) subst([D=-a/4],%),expand;
3 a2 ab a3 ac a2 b 3 a4
(%o17) [0, b − ,c − + ,d − + − ]
8 2 8 4 16 256
A UNIVERSAL METHOD OF SOLVING QUARTIC EQUATIONS 259

(%i18) subst([D=%[2],a=%[1],b=%[2],c=%[3],d=%[4]],%o12),expand;
3 a2 3 a4 a3 c a 2 b2 a4 b a6
(%o18) [2 b − , −4 d + a c + b2 − a2 b + , −c2 + a b c − − + − ]
4 16 4 4 8 64
(%i19) factor(%[3]); ´2
8 c − 4 a b + a3
`
(%o19) −
64
(%i20) y^3+%th(2)[1]*y^2+%th(2)[2]*y+%th(2)[3];
3 a2 3 a4 a 3 c a 2 b2
„ « „ «
(%o20) y3 + 2 b − y 2 + −4 d + a c + b2 − a2 b + y − c2 + a b c − − +
4 16 4 4
a4 b a6
−
8 64
(%i21) subst([y=(a^3+8*c-4*a*b)/(4*(a+4*yYFB))],%),ratsimp$
(%i22) ` -num(%)/(a^3+8*c-4*a*b),ratsimp;
8c − 4ab + a3 yY F B 3 + 16d + 2ac − 4b2 + a2 b yY F B 2 +
´ ` ´
(%o22)
` ` 2 ´ ´ 2 2
+ 8ad + a − 4b c yY F B + a d − c
(%i23) subst([y=(a^3+8*c-4*a*b)/(16*yCB)],%th(3)),ratsimp$
(%i24) expandwrt(ratsimp(num(%)/coeff(expand(num(%)),yCB,3)),
yCB)$
(%i25) create_list(coeff(%,yCB,n),n,[2,1,0]),ratsimp;
64 d − 16 a c − 16 b2 + 16 a2 b − 3 a4 8 b − 3 a2 8 c − 4 a b + a3
(%o25) [ ,− ,− ]
32 c − 16 a b + 4 a 3 16 64
260

View publication stats