Nonlinear discontinuous Petrov-Galerkin methods

Carsten Carstensen,Philipp Bringmann,Friederike Hellwig,Peter Wriggers

DOI: https://doi.org/10.48550/arXiv.1710.00529

2017-10-02

Abstract:The discontinuous Petrov-Galerkin method is a minimal residual method with broken test spaces and is introduced for a nonlinear model problem in this paper. Its lowest-order version applies to a nonlinear uniformly convex model example and is equivalently characterized as a mixed formulation, a reduced formulation, and a weighted nonlinear least-squares method. Quasi-optimal a priori and reliable and efficient a posteriori estimates are obtained for the abstract nonlinear dPG framework for the approximation of a regular solution. The variational model example allows for a built-in guaranteed error control despite inexact solve. The subtle uniqueness of discrete minimizers is monitored in numerical examples.

Numerical Analysis

What problem does this paper attempt to address?

The problems that this paper attempts to solve are related to the theory and application of the nonlinear discontinuous Petrov - Galerkin (dPG) method. Specifically, the authors introduce a least - residual method for nonlinear model problems, which uses a broken test space. The main contributions of the paper include:

1. **Method Introduction**: The paper presents the nonlinear discontinuous Petrov - Galerkin (dPG) method, which is a least - residual method used in nonlinear model problems. This method designs an optimal test space through a broken test space to achieve maximum stability.

2. **Equivalent Representations**: The paper provides several equivalent representation methods, including the mixed form, the reduced form, and the weighted nonlinear least - squares method. These representation methods provide different perspectives for understanding and analyzing the nonlinear dPG method.

3. **A Priori and A Posteriori Error Estimates**: The paper obtains quasi - optimal a priori error estimates and reliable a posteriori error estimates in the abstract nonlinear dPG framework. These estimates are crucial for evaluating the accuracy and reliability of the method.

4. **Numerical Experiments**: The paper verifies the uniqueness of the discrete minimizer through numerical experiments and demonstrates the effectiveness of the method.

5. **Model Problem**: The paper selects a specific nonlinear uniformly convex model problem as the research object, which involves energy minimization and the stress - strain relationship of Hencky materials.

### Specific Problem Description

- **Nonlinear Mapping**: The paper replaces the bilinear form \(b\) with a nonlinear mapping \(b: X\times Y\rightarrow\mathbb{R}\), where \(b\) is linear and bounded in the second component, in order to calculate the dual norm in the least - residual method.
- **Energy Minimization Problem**: The paper studies an energy minimization problem, which involves a Hilbert space setting and a quadratic nonlinearity of the gradient. The specific model problem is expressed as:
\[
f + \text{div}(\sigma(\nabla u)) = 0\quad\text{a.e. in }\Omega
\]
where \(\sigma(F)\) is a nonlinear function of the deformation gradient \(F\).
- **Discrete Stability**: The paper discusses the role of the stability condition (1.1) in the nonlinear setting, especially its application in low - order discretization.

### Main Results

- **Equivalent Representations**: The paper provides several equivalent representations of the nonlinear dPG method, including the mixed form and the weighted least - squares method.
- **Convergence Analysis**: The paper conducts the first convergence analysis in the natural norm and proves the local quasi - optimal a priori error estimates.
- **A Posteriori Error Estimates**: The paper extends the computable error estimators in linear problems and proposes a posteriori error estimates for nonlinear model problems.

Overall, this paper aims to provide a new and effective tool for the numerical solution of nonlinear partial differential equations by introducing and analyzing the nonlinear dPG method.