--- title: Cubic-Quartic Equations Overview url: https://www.emergentmind.com/topics/cubic-quartic-equation type: topic --- # Cubic-Quartic Equations Overview “Cubic-quartic equation” does not denote a single universally fixed object. In the cited literature it refers, depending on context, to classical cubic and quartic polynomial equations, to coefficient-based classification problems for real roots, to solvability criteria by radicals or by iteration, to nonlinear Schrödinger and Gross–Pitaevskii equations containing cubic and quartic terms, and to optimization subproblems formed by a cubic model with quartic regularization [2206.03855] [2002.04001] [2312.10283]. The unifying feature is the interaction between third- and fourth-order structure, but the mathematical meaning changes substantially across algebra, analysis, and numerical optimization. ## 1. Terminological scope and principal meanings In classical algebra, the relevant objects are the monic cubic \[ x^3+ax^2+bx+c \] and the monic quartic \[ x^4+ax^3+bx^2+cx+d, \] together with their depressed forms, discriminants, resolvents, and root configurations [2206.03855]. In this setting, “cubic-quartic” commonly means the joint study of cubic and quartic equations rather than a single mixed-degree scalar equation. In nonlinear-wave and condensate theory, the phrase denotes evolution equations whose nonlinear response contains both cubic and quartic terms. A representative example is the trapped cubic–quartic nonlinear Schrödinger equation \[ i\frac{\partial\psi}{\partial t} = -\mathcal{A}\frac{\partial^2\psi}{\partial x^2} +\mathcal{B}|\psi|^2\psi +\mathcal{C}|\psi|^3\psi +\mathcal{D}(x)\psi, \] where the cubic term is the mean-field Gross–Pitaevskii contribution and the quartic term is a beyond-mean-field correction associated with Lee–Huang–Yang-type physics [2002.04001]. In optimization, the expression refers neither to a scalar cubic nor to a quartic polynomial equation in the classical sense. It refers to the optimality system for minimizing a cubic multivariate model with quartic regularization, and the computational core becomes a secular equation in an auxiliary parameter \(\lambda\) [2312.10283]. A recurrent misconception is therefore that every “cubic-quartic equation” is a one-variable algebraic equation. The cited literature shows that this is false in both PDE and optimization contexts. ## 2. Algebraic equations, geometry, and coefficient structure A geometric treatment of cubic and quartic equations associates real roots of a cubic with the Siebeck–Marden–Northshield equilateral triangle and real roots of a quartic with a regular tetrahedron [2206.03855]. For the cubic \[ p_3(x)=x^3+ax^2+bx+c, \] the centroid projects to \(-a/3\), the critical points are \[ \mu_{1,2}=-\frac a3\pm \frac13\sqrt{a^2-3b}, \] and the condition \(a^2-3b>0\) is the condition that the triangle exists. The paper gives a necessary and sufficient condition for three real roots: \[ a^2-3b>0 \] together with \[ - \frac{2a^3}{27}+\frac{ab}{3}-\frac{2}{27}\sqrt{(a^2-3b)^3} \le c \le - \frac{2a^3}{27}+\frac{ab}{3}+\frac{2}{27}\sqrt{(a^2-3b)^3}. \] Under this parametrization, variation of the free term rotates the triangle and yields Viète-type trigonometric formulas for the three real roots. For the quartic \[ p_4(x)=x^4+ax^3+bx^2+cx+d, \] the tetrahedral picture places the centroid at \(-a/4\), and the inflection points are the roots of \[ p_4''(x)=12x^2+6ax+2b, \] namely \[ \rho_{1,2}=-\frac a4\pm \frac{\sqrt3}{12}\sqrt{3a^2-8b}. \] The quartic discriminant is cubic in \(d\), so varying the free term again becomes structurally important. The condition for four distinct real roots is expressed as \[ b<\frac38 a^2,\qquad C_20 \Rightarrow 3 \text{ distinct real roots},\qquad D<0 \Rightarrow 1 \text{ real root and 2 complex conjugates}. \] When \(D=0\), the condition \(p^2-3q\neq 0\) gives one double real root and one simple real root, while \(p^2-3q=0\) gives a triple root. The double root is \[ \frac{-9r+pq}{2(p^2-3q)}, \] and the sign of \[ 27r-9pq+2p^3 \] determines whether the simple root lies to its left or right. For the quartic \[ x^4+px^3+qx^2+rx+s, \] the discriminant \(D\) is supplemented by \[ L_1=p^2q^2-3p^3r-6p^2s-4q^3+14pqr+16qs-18r^2 \] and \[ L_2=3p^2-8q. \] The classification includes: \[ D>0,\ L_2>0,\ L_1>0 \Rightarrow 4 \text{ distinct real roots}, \] \[ D<0 \Rightarrow 2 \text{ real roots and 2 complex conjugates}, \] \[ D=0,\ L_1>0 \Rightarrow 1 \text{ real double root and 2 real simple roots}, \] \[ D=0,\ L_1<0 \Rightarrow 1 \text{ real double root and 2 complex conjugates}, \] \[ D=0,\ L_1=0,\ 3p^2-8q>0,\ p^3-4pq+8r=0 \Rightarrow 2 \text{ real double roots}, \] \[ D=0,\ L_1=0,\ 3p^2-8q>0,\ p^3-4pq+8r\neq 0 \Rightarrow 1 \text{ triple real root and 1 simple real root}, \] and \[ D=0,\ L_1=0,\ 3p^2-8q=0 \Rightarrow \text{quadruple root}. \] A different but related line of work studies positivity instead of mere root count. For the cubic \[ h(t)=t^3+pt^2+qt+r,\qquad r>0, \] strict positivity on \(t\ge 0\) is equivalent to either \[ p\ge 0,\ q\ge 0 \] or \[ \Delta(h)<0, \] where \[ \Delta(h)=p^2q^2+18pqr-27r^2-4p^3r-4q^3. \] The boundary case \(\Delta(h)=0\) with either \(p<0\) or \(q<0\) gives nonnegativity but not strict positivity, and the unique positive zero is \[ \alpha=\frac{9r-pq}{2p^2-6q}>0 \] [2008.10922]. For the normalized quartic \[ f(t)=t^4+\alpha t^3+\beta t^2+\gamma t+1, \] nonnegativity on all real numbers is equivalent to \[ \Delta(f)\ge 0,\qquad |\alpha-\gamma|\le 4\sqrt{\beta+2}, \] and either \[ -2\le \beta\le 6 \] or \[ \beta>6,\qquad |\alpha+\gamma|\le 4\sqrt{\beta-2}. \] Strict positivity allows either the same inequalities with \(\Delta(f)>0\), or the discriminant-zero “appendix curve” \[ \alpha=\gamma,\qquad \alpha^2+8=4\beta<24. \] A notable consequence is that \(\Delta(f)=0\) need not indicate a boundary point of the quartic nonnegativity region; on the appendix curve it lies inside the interior [2008.10922]. ## 4. Solvability, radicals, and iterative cubic computation Over \(\mathbb{Q}\), 1-solvability is stricter than ordinary solvability by radicals. For the depressed cubic \[ x^3+px+q, \] an irreducible cubic is 1-solvable if and only if it has a rational root or \[ D_{pq}=\left(\frac p3\right)^3+\left(\frac q2\right)^2\ge 0 \] with \[ \sqrt{D_{pq}}\in\mathbb{Q}. \] For the irreducible depressed quartic \[ x^4+px^2+qx+s,\qquad q\ne 0, \] 1-solvability is equivalent to the existence of a rational root \(t\) of the cubic resolvent \[ q^2-4(p-2t)(s-t^2)=0 \] such that \(t>p/2\) and \[ A=16(t^2-s)^2-(t^2-s)(2t+p)^2 \] is a square of a rational number [1411.4990]. This places the quartic criterion explicitly under the control of an auxiliary cubic. A structurally different derivation of cubic radicals uses Harrison’s center theory. For \[ f(x)=a_0x^3+3a_1x^2+3a_2x+a_3, \] the associated binary cubic can be diagonalized into a sum of two cubes of linear forms, yielding a completed-cube decomposition and, in the depressed case, Cardano’s formula. The same paper extends the “completing powers” philosophy to special higher-degree equations, but explicitly states that even for quartics the method works only for very special equations [2301.07375]. The literature also contains iterative alternatives that avoid using Cardano as the computational mechanism. One geometric-iterative method starts from a monic cubic complex polynomial with distinct roots and distinct critical points, proves that at least one critical point has the Voronoi property, and then applies Kalantari’s basic family \[ B_m(\xi)=\xi-p(\xi)\frac{d_{m-2}}{d_{m-1}} \] to converge to a root. The quadratic formula enters only through the computation of the critical points of \(p'(z)=0\), so the method is “solving a cubic by the quadratic formula” in an iterative rather than radical sense [1401.5148]. Another numerical program, inspired by Sharaf al-Din Tusi and Smale’s point-estimation theory, reduces any real cubic to one of four canonical forms and then chooses a certified Newton seed from \[ \{\rho_q,\ \pm 0.95\rho_q,\ -1/3,\ 1\}, \] where \(\rho_q\) approximates \(\sqrt[3]{q}\) within five percent relative error. For suitable canonical forms, after \(t\) Newton iterations the error satisfies \[ |x_t-\theta_q|\le \sqrt[3]{q}\cdot 2^{-2^t}, \] providing a high-precision route to a real root without Cardano’s formula [2303.17747]. ## 5. Cubic-quartic nonlinear Schrödinger and Gross–Pitaevskii equations In ultracold-gas theory, the cubic–quartic nonlinear Schrödinger equation models the competition between mean-field and beyond-mean-field interactions. In quasi-1D, the stationary reduction of \[ i\psi_t = -\mathcal{A}\psi_{xx} +\mathcal{B}|\psi|^2\psi +\mathcal{C}|\psi|^3\psi +\mathcal{D}(x)\psi \] leads to \[ \phi''+\alpha\phi-\beta|\phi|^2\phi-\gamma|\phi|^3\phi-\delta(x)\phi=0. \] With a matched cnoidal trap, \[ \delta(z)=V_0\,\mathrm{cn}^3(z,m) \quad \text{or} \quad \delta(z)=V_0\,\mathrm{sn}^3(z,m), \] the equation admits exact \(\mathrm{cn}\)- and \(\mathrm{sn}\)-families. The \(\mathrm{cn}\)-branch yields, in the limit \(m\to 1\), the localized solution \[ \phi(z)=-\frac{\beta}{2\gamma}\left(1\mp \sqrt{2}\,\mathrm{sech}\,z\right), \] while the \(\mathrm{sn}\)-branch yields the delocalized kink–antikink form \[ \phi(z)=-\frac{\beta}{2\gamma}\left(1\mp \frac{1}{\sqrt2}\tanh^{-1} z\right) \] in the sense stated in the source formulas. A central conclusion is the absence of a nontrivial sinusoidal mode: in both branches the \(m\to 0\) limit forces the oscillatory amplitude to zero [2002.04001]. A driven version in a bi-chromatic optical lattice adds a phase-locked source term and admits the exact periodic solution \[ \psi(x)= -\frac{g_1}{3g_2}+\left(\frac{V_1}{g_2}\right)^{1/3}\sin(\zeta x), \] with \[ V_2=\frac{Bg_1^2}{3g_2},\qquad \mu=\frac{2g_1^3}{27g_2^2},\qquad \zeta=\pm\left(\frac{2F}{B}\right)^{1/2}. \] Its density forms a stripe or density-wave pattern, and VK analysis gives stability for \(g_1>0\). The same work extends the model to quadratic-cubic-quartic and quadratic-cubic nonlinearities in order to describe dimensional crossover between quasi-1D and strict 1D [2107.11779]. A nonlocal cubic–quartic Gross–Pitaevskii equation in \(\mathbb{R}^3\), \[ i\,\partial_t \psi = -\frac{1}{2}\Delta \psi +\lambda_1|\psi|^2\psi +\lambda_2(K*|\psi|^2)\psi +\lambda_3|\psi|^3\psi, \qquad \lambda_3<0, \] is treated variationally on the mass shell \(\|\psi\|_2^2=c\). Because the constrained energy is unbounded below, ground states are constructed as mountain-pass critical points on the \(L^2\)-sphere. The analysis uses localized Palais–Smale sequences and the nonlocal Pohozaev identity \[ Q(u)=A(u)+\frac32 B(u)+\frac95 C(u)=0, \] which has no remainder term for the dipolar kernel. For suitable masses, ground states exist, are positive up to a constant phase, and have positive chemical potential [1806.00697]. Another integrable usage occurs in an extended nonlinear Schrödinger equation with third- and fourth-order corrections, \[ iq_t+q_{xx}+2|q|^2q -i\alpha(q_{xxx}+6q_x|q|^2) +\gamma(\cdots +6|q|^4q)=0. \] Although the title speaks of cubic and quartic nonlinearity, the full model also contains derivative nonlinear couplings and the quintic term \(6|q|^4q\). Using the generalized Darboux transformation, the paper constructs higher-order smooth positons and breather-positons and derives the exact time-dependent displacement \[ \Delta(t)= \frac{1}{a}\log\!\bigl[4 a^2 |t| \sqrt{\hat \kappa}\bigr], \] showing direct dependence on the higher-order parameters \(\alpha\) and \(\gamma\) [2207.12095]. This is another context in which “cubic-quartic” is descriptive but not exhaustive. ## 6. Cubic-quartic regularization in higher-order optimization In third-order tensor methods for optimization, the relevant model is the quartically regularized cubic subproblem \[ m_3(s) = f_0+g^Ts+\frac12 H[s]^2+\frac16 T[s]^3+\frac{\sigma}{4}\|s\|^4. \] The proposed CQR framework replaces the exact cubic tensor term by a simpler local model \[ M_c(s) = f_i+g_i^Ts+\frac12 H_i[s]^2+\frac{\beta}{6}\|s\|_W^3+\frac{\sigma_c}{4}\|s\|_W^4, \] where the cubic term approximates local tensor information and the quartic term regularizes progress [2312.10283]. The global minimizer \(s_c\) is characterized by \[ B(s_c)s_c=-g_i, \qquad B(s_c)\succeq 0, \] with \[ B(s_c)=H_i+\frac{\beta}{2}W\|s_c\|_W+\sigma_c W\|s_c\|_W^2. \] Introducing \[ \lambda=\frac{\beta}{2}\|s(\lambda)\|_W+\sigma_c\|s(\lambda)\|_W^2 \] reduces the optimality system to \[ (H_i+\lambda W)s(\lambda)=-g_i,\qquad H_i+\lambda W\succeq 0, \] together with the scalar secular equation \[ \psi(\lambda)=k_\pm(\lambda). \] Accordingly, the “cubic-quartic equation” here is a coupled optimality system plus scalar root-finding reduction, not a classical scalar quartic equation. The paper proves necessary and sufficient optimality conditions for global minimizers and shows that a CQR variant achieves the optimal-order evaluation complexity \[ \mathcal{O}(\epsilon^{-3/2}) \] for minimizing the quartically regularized cubic subproblem, with improvements in special cases. In the univariate case, the quartically regularized cubic can be solved within a single CQR iteration by minimizing two scalar cubic-quartic models, and numerically the method is reported to be competitive with ARC and QQR, with particularly strong performance on ill-conditioned and badly scaled instances [2312.10283]. Across these literatures, the cubic–quartic theme is structurally consistent even when the objects differ. In algebra it organizes discriminants, resolvents, and geometric root models; in analysis it encodes the competition of mean-field and higher-order effects; and in optimization it links higher-order local models to secular equations and regularization. The term is therefore best understood not as a single definition, but as a family of mathematically precise settings in which cubic and quartic mechanisms interact. Source: https://www.emergentmind.com/topics/cubic-quartic-equation