Existence results for a Cauchy-Dirichlet parabolic problem with a repulsive gradient term
Martina Magliocca
TL;DR
This work analyzes existence for a nonlinear parabolic Cauchy-Dirichlet problem with a gradient-dependent lower-order term H(t,x,∇u) that grows like γ|∇u|^q plus a forcing f. By exploiting Leray–Lions type operators and a careful distinction between sublinear, linear, and superlinear gradient growth, the authors establish finite-energy and renormalized-solutions theories under sharp relations between q, p, N, and the data spaces of the initial datum and forcing. The paper derives extensive a priori estimates, compactness, and convergence results for approximating problems, enabling existence results across multiple regimes (finite-energy, infinite-energy, and L^1 data). It thus extends parabolic gradient-growth analysis to general divergence-form operators with measurable coefficients, bridging gaps between sublinear growth and natural-growth settings. The results have implications for well-posedness of parabolic PDEs with gradient nonlinearities in heterogeneous media and contribute a robust framework for handling unbounded data and gradient-driven source terms.
Abstract
We study the existence of solutions of a nonlinear parabolic problem of Cauchy-Dirichlet type having a lower order term which depends on the gradient. The model we have in mind is the following: \[ \begin{cases}\begin{split} & u_t-\text{div}(A(t,x)\nabla u|\nabla u|^{p-2})=γ|\nabla u|^q+f(t,x) &\qquad\text{in } Q_T,\\ & u=0 &\qquad\text{on }(0,T)\times \partial Ω,\\ & u(0,x)=u_0(x) &\qquad\text{in } Ω, \end{split}\end{cases} \] where $Q_T=(0,T)\times Ω$, $Ω$ is a bounded domain of $\mathrm{R}^N$, $N\ge 2$, $1<p<N$, the matrix $A(t,x)$ is coercive and with measurable bounded coefficients, the r.h.s. growth rate satisfies the superlinearity condition \[ \max\left\{\frac{p}{2},\frac{p(N+1)-N}{N+2}\right\}<q<p \] and the initial datum $u_0$ is an unbounded function belonging to a suitable Lebesgue space $L^σ(Ω)$. We point out that, once we have fixed $q$, there exists a link between this growth rate and exponent $σ=σ(q,N,p)$ which allows one to have (or not) an existence result. Moreover, the value of $q$ deeply influences the notion of solution we can ask for. The sublinear growth case with \[ 0<q\le\frac{p}{2} \] is dealt at the end of the paper for what concerns small value of $p$, namely $1<p<2$.
