Abstract:We develop a general toolbox to study $W^{1,p}$ solutions of differential inclusions $\nabla u \in K$ for unbounded sets $K$. A key notion is the concept that a subset $K$ of the space $\mathbb{R}^{d \times m}$ of $d \times m$ matrices can be reduced to another set $K'$. We then use this framework to show that the product rigidity for Sobolev maps fails for $p<2$, and also apply our toolbox to simplify several examples from the literature.
What problem does this paper attempt to address?
This paper attempts to address two main issues:
1. **Verify whether the rigidity results for Sobolev mappings fail at low Sobolev indices**:
- The paper first demonstrates that the rigidity results for mappings \( f: \Omega_1 \times \Omega_2 \subset \mathbb{R}^n \times \mathbb{R}^n \to \mathbb{R}^n \times \mathbb{R}^n \) in the Sobolev space \( W^{1,p} \) from the previous paper [KMSJX23] are sharp (i.e., optimal) when \( p < 2 \). Specifically, for all \( n \geq 2 \) and \( p < 2 \), there exists a mapping \( f \in W^{1,p} \) such that its weak derivative \( \nabla f \) is almost everywhere invertible and preserves or reverses the product structure, but \( f \) does not have the form \( f(x, y) = (f_1(x), f_2(y)) \) or \( f(x, y) = (f_2(y), f_1(x)) \).
2. **Develop a general toolbox for studying differential inclusion problems of Sobolev solutions**:
- Secondly, the paper develops a general toolbox for studying \( W^{1,p} \) solutions to the differential inclusion problem \( \nabla u \in K \), where \( K \) is an unbounded set. A key concept is the notion that the set \( K \) can be reduced to another set \( K' \) (see Definition 1.1). Through this approach, optimal \( L^p \) regularity and irregular solutions for elliptic systems and \( p \)-Laplace equations can be easily obtained.
### Main Results
- **Theorem 1.2**: It is proven that when \( p < 2 \), there exists a mapping \( u \in W^{1,1}(\Omega) \cap C^\alpha(\Omega) \) that satisfies specific conditions, particularly \( u \in W^{1,p}(\Omega) \) for any \( p < 2 \), and if \( A \notin L \), then \( u \) is not split.
- **Theorem 1.4**: It is proven that when \( p < 2n \), there exists a sequence of mappings \( u(j) \) that satisfy specific conditions, particularly \( u(j) \) weakly converges to a linear mapping \( l_A \) in \( W^{1,p}(\Omega) \), but the distance of \( \nabla u(j) \) to \( L \cap \Sigma \) does not tend to zero in the \( L^1 \) norm.
### Proof Strategy
- **Phase Construction**:
- **Phase 1**: Reduce \( \mathbb{R}^{2n \times 2n} \) to the set of matrices with rank at most 1, in weak \( L^2 \).
- **Phase 2**: Reduce the set of matrices with rank at most 1 to the set \( L \), in weak \( L^q \) for all \( q < \infty \).
- **Phase 3**: Reduce \( L_i \) to the set \( L_i \cap \Sigma \), in weak \( L^{2n} \) for \( i = 1, 2 \).
### Key Concepts
- **Staircase Laminates**: A special type of probability measure used to construct mappings that satisfy specific conditions. By constructing step by step, the mass can be "pushed to infinity," thereby achieving the required weak \( L^p \) estimates.
Through these methods, the paper not only addresses the aforementioned issues but also provides powerful tools for studying other related problems.