Answers are grounded in this paper and cite their sources.
Purpose This study aims to develop and assess state-dependent nonlinear eigenvalue-problem formulations within the generalized integral transform technique (GITT) for transient nonlinear diffusion problems. Design/methodology/approach A hybrid numerical-analytical solution for nonlinear diffusion is formulated by incorporating the nonlinear equation coefficients directly into the eigenvalue problem used to construct the expansion, instead of transferring them to source terms as in the classical GITT approach. The nonlinear eigenfunctions are expanded through an auxiliary linear Sturm–Liouville basis, reducing the nonlinear differential eigenvalue problem to a nonlinear algebraic one. The resulting transformed system couples the nonlinear potentials, eigenvalues and eigenvectors. A variant approach is also proposed that incorporates the non-transformed term of the partial differential equation (PDE) back into the nonlinear eigenvalue problem, leading to a decoupled ordinary differential equations (ODEs) system with an analytical solution. Findings Monotone convergence of the maximum absolute error was observed for the nonlinear-basis formulation across all cases, nonlinearity strengths and time instants examined, while the classical linear-eigenbasis GITT implementation showed non-monotone convergence and increasingly pronounced error plateaus as the nonlinearity intensified. However, this improved convergence predictability was obtained at a substantially higher computational cost because of the enlarged differential-algebraic system. The variant of the state-dependent eigenvalue problem is even faster converging than the first considered nonlinear eigenvalue problem, also showing a similar monotonic decaying pattern, which leads to computational savings for a comparable accuracy level. Originality/value This work introduces a state-dependent nonlinear eigenvalue-problem basis within the GITT framework, providing a more self-consistent treatment of nonlinear operators than conventional linear-basis formulations. The proposed strategy is most advantageous for strongly nonlinear regimes, high-accuracy requirements, or situations where predictable convergence is essential, while the classical formulation remains more efficient for moderate nonlinearities.
No takes yet. Share an insight, caveat, or question.
Lisbôa et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: