Adjoint-based anisotropic hp-adaptation for discontinuous Galerkin methods using a continuous mesh model

In this paper we propose an adjoint-based hp-adaptation method for conservation laws, and corresponding numerical schemes based on piecewise polynomial approximation spaces. The method uses a continuous mesh framework, similar to that proposed in [1], where a global optimization scheme was formulated with respect to the error in the numerical solution, measured in any Lq norm. The novelty of the present work is the extension to more general optimization targets. Here, any solution-dependent functional, which is compatible with an adjoint equation, may be the target of the continuous-mesh optimization. We present the rationale behind the formulation of the optimization problem, with particular emphasis on the continuous mesh model, and the relevant adjoint-based error estimate. Additionally we combine the adjoint-based error estimates with the polynomial optimization strategy from [2] to present a complete hp-adaptation method which shows exponential convergence in the target function. The h-only mesh adaptation strategy of this work has been presented as a conference proceeding earlier [3]. Numerical experiments are carried out to demonstrate the viability of the scheme.

Download paper here

Recommended citation: Ajay Rangarajan, Georg May, Vit Dolejsi. (2020). "Adjoint-based anisotropic hp-adaptation for discontinuous Galerkin methods using a continuous mesh model." Journal of Computational Physics