--- title: Generalized Burgers–Fisher Equation url: https://www.emergentmind.com/topics/generalized-burgers-fisher-equation type: topic --- # Generalized Burgers–Fisher Equation The generalized Burgers–Fisher equation denotes a family of one-dimensional nonlinear advection–diffusion–reaction equations that combine Burgers-type convection, viscous diffusion, and Fisher-type reaction. A common constant-coefficient form is \[ u_t+\beta\,u^\delta u_x-u_{xx}=\gamma\,u(1-u^\delta), \] while equivalent parameterizations write \[ U_t+\sigma_1\,U^\delta U_x-\mu\,U_{xx}=\sigma_2\,U(1-U^\delta). \] Time-fractional extensions replace the first-order time derivative by a derivative of order \(0<\alpha\le 1\), and more recent dynamical-systems work studies the closely related generalized Burgers–Fisher–KPP law \(\partial_tu=u_{xx}+k(u^n)_x+u^p-u^q\) [2008.01695, 2306.09988, 2003.05295, 2509.24909]. ## 1. Canonical forms and parameter regimes In the constant-coefficient setting, the generalized Burgers–Fisher equation is usually written as \[ u_t+\beta\,u^\delta u_x-u_{xx}=\gamma\,u(1-u^\delta), \] where \(u=u(x,t)\), \(\beta\) weights convection, the term \(-u_{xx}\) provides diffusion, and \(\gamma\,u(1-u^\delta)\) is a Fisher-type reaction. The exponent \(\delta>0\) controls the nonlinearity in both the convective and reactive contributions [2008.01695]. A numerically oriented parameterization replaces \((\beta,\gamma)\) by \((\sigma_1,\sigma_2)\) and allows an explicit diffusion coefficient \(\mu>0\), \[ U_t+\sigma_1\,U^\delta U_x-\mu\,U_{xx}=\sigma_2\,U(1-U^\delta), \] typically posed on a rectangular cylinder with prescribed initial and boundary data; in that formulation \(\sigma_1\ge 0\), \(\sigma_2\ge 0\), \(\mu>0\), and \(\delta>0\) [2306.09988]. Several papers extend this baseline model in structurally distinct directions. Selvaraj et al. consider time-fractional generalized Burgers–Fisher equations with either the Caputo derivative \[ {}^C D_t^\alpha u + B\,u\,u_x-u_{xx}=y\,u(1-u^\delta), \qquad 0<\alpha\le 1, \] or, in a companion treatment, the same Caputo-based model analyzed by Lie symmetry methods [2003.05295, 2003.05294]. A different time-fractional study uses Jumarie’s modified Riemann–Liouville derivative and writes \[ D_t^\alpha u+\beta\,u^\delta u_x-u_{xx}=\gamma\,u(1-u^\delta), \qquad 0<\alpha\le 1 \] [2008.01695]. Variable-coefficient work replaces constants by time-dependent functions, \[ u_t-u_{xx}+\alpha(t)\,u\,u_x=\beta(t)\,u(1-u), \] thereby coupling Burgers–Fisher dynamics to explicitly nonautonomous coefficients [1004.1815]. A further extension broadens both the convective and reactive laws. Iagar–Sánchez study \[ \partial_tu=u_{xx}+k(u^n)_x+u^p-u^q,\qquad n\ge 2,\quad p>q\ge 1,\quad k>0, \] and the subsequent large-time analysis of the same equation treats \(k\in\mathbb R\) and identifies two distinct critical velocities for Heaviside and anti-Heaviside data [2509.24909, 2604.22108]. Garrione and Strani generalize the diffusion itself through a saturating flux, \[ u_t=\partial_x\!\bigl[P(\partial_x(D(u)))\bigr]-\partial_x(h(u))+f(u), \] with \(P(s)=s/\sqrt{1+s^2}\) as the model mean-curvature case, density-dependent diffusion \(D\), and a general Fisher–Burgers reaction term \(f\) [1702.03782]. This range of formulations shows that the name “generalized Burgers–Fisher equation” is used for a class of related advection–diffusion–reaction models rather than a single normalized PDE. ## 2. Traveling-wave reductions and exact solution families The dominant analytic reduction is the traveling-wave ansatz. For the time-fractional equation with modified Riemann–Liouville derivative, one sets \[ u(x,t)=u(\xi),\qquad \xi=kx-\frac{\lambda t^\alpha}{\Gamma(1+\alpha)}, \] which converts the PDE into \[ k^2u''+\bigl(\lambda-k\beta u^\delta\bigr)u'+\gamma\,u(1-u^\delta)=0. \] A standard folding transformation, \[ u(\xi)=\bigl[v(\xi)\bigr]^{1/\delta}, \] then yields a polynomially structured nonlinear ODE for \(v\), more amenable to algebraic ansatz methods [2008.01695]. Selvaraj et al. perform an analogous reduction for the Caputo-based model using \(\xi=kx-\lambda t^\alpha\), obtaining a nonlinear ODE and then a “master” second-order equation after the same type of folding transformation [2003.05295]. The generalized Kudryashov method introduces an auxiliary Riccati-type equation, \[ R'=R^2-R, \] and assumes a rational-polynomial representation \[ V(\xi)=\frac{p_0+p_1R+p_2R^2+p_3R^3}{q_0+q_1R}. \] Homogeneous balance fixes the degrees \(K=3\) and \(N=1\), and coefficient matching produces exact wave families. Representative solutions include the kink-type fronts \[ u_1(x,t)=\Bigl[1+e^{-(kx-\lambda t^\alpha)}\Bigr]^{-1/\delta},\qquad u_2(x,t)=\Bigl[1+e^{\,kx-\lambda t^\alpha}\Bigr]^{-1/\delta}, \] and the hyperbolic-function family \[ u_3(x,t)=\bigl[1+D\,\sech(kx-\lambda t^\alpha)\bigr]^{1/\delta},\qquad u_4(x,t)=\bigl[1+D\,\csch(kx-\lambda t^\alpha)\bigr]^{1/\delta}, \] with \(D\) and \(\lambda\) determined algebraically from the coefficient system [2003.05295]. The Exp-function method and the Exponential Rational Function Method produce a broader catalog of exact traveling waves. In the Exp-function approach, the folded variable is approximated by \[ v(\xi)=\frac{\sum_{q=-g}^{h}a_q e^{q\xi}}{\sum_{p=-k}^{j}b_p e^{p\xi}}, \] and balancing yields the minimal nontrivial choice \[ v(\xi)=\frac{a_1e^\xi+a_0+a_{-1}e^{-\xi}}{b_1e^\xi+b_0+b_{-1}e^{-\xi}}. \] For the time-fractional generalized Burgers–Fisher equation, this procedure gives nine families of traveling-wave solutions. The Exponential Rational Function Method instead assumes \[ v(\xi)=A_0+\frac{A_1}{1+e^\xi}+\frac{A_2}{(1+e^\xi)^2}, \] and yields seven additional families [2008.01695]. The reported solution structures are not limited to monotone fronts. The exact-solution literature explicitly identifies monotonic fronts, kink or shock profiles, bell-shaped solitary waves, localized pulses, and hyperbolic-cosine forms, with the shape controlled by the algebraic relations among \(\beta\), \(\gamma\), \(\delta\), \(k\), and \(\lambda\) [2008.01695]. In the Caputo-based Kudryashov analysis, the limit \(\alpha\to 1\) recovers the known integer-order traveling waves, while for \(0<\alpha<1\) the phase variable \(kx-\lambda t^\alpha\) produces fractional self-similar wave profiles [2003.05295]. ## 3. Fractional formulations, Lie symmetries, and series representations A distinguishing feature of the fractional literature is that different derivative definitions lead to different reduction formulas. One line of work uses the Caputo derivative, \[ {}^C D_t^\alpha u(x,t)=\frac1{\Gamma(1-\alpha)}\int_0^t (t-s)^{-\alpha}\,\frac{\partial u}{\partial s}(x,s)\,ds, \] while another uses Jumarie’s modified Riemann–Liouville derivative. Both formulations treat \(0<\alpha\le 1\) as the fractional order and reinterpret the equation as a reaction–diffusion–convection model with time memory [2003.05295, 2008.01695]. Selvaraj et al. derive Lie point symmetries for the Caputo-based time-fractional generalized Burgers–Fisher equation and find a two-dimensional Lie algebra generated by \[ X_1=\partial_x,\qquad X_2=2t\,\partial_t+\alpha x\,\partial_x-\alpha u\,\partial_u, \] with commutator \([X_1,X_2]=\alpha X_1\) [2003.05294]. The scaling generator \(X_2\) yields the invariants \[ \zeta=x\,t^{-\alpha/2},\qquad u(x,t)=t^{-\alpha/2}f(\zeta), \] which reduce the PDE to a nonlinear fractional ODE involving the Erdélyi–Kober operator, \[ P_{\,2-\tfrac{\alpha}{2},\,\alpha}^{(\zeta)}[f(\zeta)] +B\,f(\zeta)f'(\zeta)-f''(\zeta)-\gamma f(\zeta)\bigl(1-f(\zeta)^\delta\bigr)=0. \] This is a similarity reduction rather than a traveling-wave reduction, and it isolates the self-similar scaling imposed by the fractional symmetry group [2003.05294]. The same paper develops a power-series solution \[ f(\zeta)=\sum_{n=0}^\infty b_n\zeta^n, \] with two free parameters \(b_0\) and \(b_1\), and derives a recurrence for \(b_{n+2}\) involving gamma-function coefficients and nonlinear convolution sums [2003.05294]. In the original variables, the resulting exact series solution is \[ u(x,t)=t^{-\alpha/2}\sum_{n=0}^\infty b_n\bigl[x\,t^{-\alpha/2}\bigr]^n. \] The integer-order limit \(\alpha\to 1\) reduces the Caputo derivative to \(\partial_t\) and the Erdélyi–Kober operator to an ordinary derivative, thereby recovering the classical similarity ODE [2003.05294]. Across the exact-solution literature, the fractional order is consistently interpreted as modifying propagation and relaxation. The Caputo-based Kudryashov study states that the fractional order slows the approach to steady-state fronts, and the symmetry study similarly links \(\alpha<1\) to slower front propagation and delayed shock formation relative to the classical case [2003.05295, 2003.05294]. The modified Riemann–Liouville study connects variations in \(\alpha\) to changes in wave speed and amplitude visible in its 2D and 3D plots [2008.01695]. ## 4. Numerical discretization and stability theory The numerical literature represented here emphasizes high-order collocation. The time–space Chebyshev pseudospectral method (TS-CPsM) maps the initial-boundary value problem to \([-1,1]\times[-1,1]\), homogenizes the non-homogeneous data by a lifting function \(\Omega\), and approximates the unknown homogeneous component \(V\) by a double Chebyshev–Lagrange expansion on Chebyshev–Gauss–Lobatto points [2306.09988]. In contrast, the extended cubic B-spline method uses a spline basis with an extension parameter \(\lambda\), semi-discretizes in space, and advances in time by Crank–Nicolson with Rubin–Graves linearization [1612.03333]. | Method | Core construction | Reported outcome | | --- | --- | --- | | TS-CPsM | CGL collocation in time and space; lifting \(\Omega\); Newton–Raphson with assembled Jacobian | Unconditionally energy-stable; \(L_\infty\) errors from \(1.2\cdot10^{-4}\) to \(4.5\cdot10^{-17}\) as \(N\) goes from \(4\) to \(24\) for \(\delta=2,\ T=1\) | | Extended cubic B-spline collocation | Extended cubic B-splines; Crank–Nicolson; Rubin–Graves linearization; Thomas-type solver | Example 1 errors \(\sim 10^{-12}\)–\(10^{-11}\) | For TS-CPsM, the fully discrete nonlinear system is written in Kronecker-product form as \(G(\mathbf V)=0\) and solved by Newton’s method. The paper derives a discrete energy estimate \[ \frac{dE}{dt}\le 2CE(t)\qquad\Longrightarrow\qquad E(t)\le E(0)e^{2Ct}, \] where \(E(t)\) is the CGL-weighted discrete \(L^2\)-type energy. This establishes unconditional energy stability in the norm induced by CGL quadrature [2306.09988]. The reported benchmark shows spectral convergence in \(N\), and the method is described as remaining robust up to \(\delta=8\) or higher; comparisons with BSQI and one-dimensional SCM indicate that fewer time grid points are required to achieve the same accuracy [2306.09988]. The extended cubic B-spline method approximates the solution by \[ U(x,t)=\sum_{i=-1}^{N+1}E_i(x)\,\delta_i(t), \] where \(E_i\) are extended cubic B-splines of degree four. Collocation at the nodal points produces a semi-discrete ODE system, and the time-stepping scheme leads to a tridiagonal linear system after eliminating the boundary coefficients [1612.03333]. In the example with exact solution and parameters \(a=0.1\), \(\mu=-0.0025\), \(\nu=1\), \(q=1\), \(\Delta t=10^{-4}\), and \(N=16\), the maximum norm error at \(t=0.1,0.2,\dots,0.5\) is reported to be on the order of \(10^{-12}\) [1612.03333]. ## 5. Phase-plane classification and wave-speed selection For the related generalized Burgers–Fisher–KPP equation \[ \partial_tu=u_{xx}+k(u^n)_x+u^p-u^q, \] Iagar–Sánchez derive a complete phase-plane classification of traveling waves. With the ansatz \(u(x,t)=f(\xi)\), \(\xi=x+ct\), the profile satisfies \[ f''+c f'+kn f^{n-1}f'+f^p-f^q=0. \] Writing \(X=f\) and \(Y=f'\) gives the planar system \[ X'=Y,\qquad Y'=cY-knX^{n-1}Y-X^p+X^q, \] whose equilibria in the half-plane \(X\ge 0\) are \(P_1=(0,0)\) and \(P_2=(1,0)\) [2509.24909]. The local dynamics at \(P_2\) are governed by the trace \(c-kn\) and determinant \(p-q\). In particular, \(P_2\) is an unstable node for \(c>kn+2\sqrt{p-q}\), an unstable focus for \(knp+q+1\) [2509.24909]. The central global result is the existence of a unique \(c^*\in(0,\infty)\) such that there exists a unique soliton \(f^*\) with speed \(c^*\) satisfying \[ \lim_{\xi\to-\infty}f^*(\xi)=\lim_{\xi\to\infty}f^*(\xi)=0. \] Moreover, if \(np+q+1\) then \(c^*>kn\). For \(c<\min\{c^*,kn\}\), the traveling wave connects \(0\) at \(-\infty\) to \(1\) at \(+\infty\); for \(c>\max\{c^*,kn\}\), it connects \(1\) at \(-\infty\) to \(0\) at \(+\infty\) [2509.24909]. The same paper emphasizes a point absent in the non-convective case \(k=0\): for any \(c\in(0,c^*)\), there are traveling waves with speed \(c\) such that \(u(x,t)\to 1\) as \(t\to\infty\) for fixed \(x\) [2509.24909]. The large-time analysis of the Cauchy problem identifies two different selected speeds. For anti-Heaviside data, the selected velocity is explicit, \[ \widetilde c=kn+2\sqrt{p-q}, \] and the solution converges in shape to the unique heteroclinic wave with that speed. For Heaviside data, the selected speed is the anomalous velocity \[ \overline c:=\sup\{c\in\mathbb R:\text{ the trajectory }l_c\text{ connects directly }P_1\to P_2\text{ with }Y(\xi)>0\}, \] which has no closed-form algebraic expression in \((n,p,q,k)\) [2604.22108]. The sign of \(\overline c\) determines whether the Heaviside solution vanishes to \(0\) or spreads to \(1\), and there exists a threshold coefficient \(k^*(n,p,q)\) such that \(\overline c>0\) for \(k>k^*(n,p,q)\) and \(\overline c<0\) for \(k