Abstract:Given an integer or a non-negative integer solution $x$ to a system $Ax = b$, where the number of non-zero components of $x$ is at most $n$. This paper addresses the following question: How closely can we approximate $b$ with $Ay$, where $y$ is an integer or non-negative integer solution constrained to have at most $k$ non-zero components with $k<n$? We establish upper and lower bounds for this question in general. In specific cases, these bounds match. The key finding is that the quality of the approximation increases exponentially as $k$ goes to $n$.
What problem does this paper attempt to address?
### The problems the paper attempts to solve
The paper aims to study, given an integer or non - negative integer solution \(x\) that satisfies the linear equation system \(Ax = b\), and the number of non - zero components in \(x\) is at most \(n\), how to approximate \(b\) with another integer or non - negative integer solution \(y\), where the number of non - zero components of \(y\) is limited to at most \(k\) (\(k < n\)). Specifically, the author hopes to determine the best possible bounds for the following problem:
\[
\text{app}_{X,k,\|\cdot\|}(A):=\max_{b\in A\cdot X}\min\{\|Ax - b\|:x\in X,|supp(x)|\leq k\}
\]
where:
- \(A\in\mathbb{Z}^{m\times n}\) is a full - row - rank matrix.
- \(X\subseteq\mathbb{R}^n\), mainly considering \(X\in\{\mathbb{Z}^n,\mathbb{Z}^n_{\geq0}\}\).
- \(\|\cdot\|\) is a norm on \(\mathbb{R}^m\).
- \(supp(x): = \{i\in[n]:x_i\neq0\}\) represents the support set of \(x\).
### Key problems and results
1. **Upper and lower bounds**: The paper establishes general upper and lower bounds for the above approximation problem. In some specific cases, these upper and lower bounds match.
2. **Sparse approximation quality**: As \(k\) approaches \(n\), the approximation quality improves exponentially.
3. **Analysis of special cases**:
- When \(X = \mathbb{Z}^n\), if \(k\geq m+\log_2(\delta(A))\), then \(\text{app}_{\mathbb{Z}^n,k,\|\cdot\|_\infty}(A) = 0\), where \(\delta(A):=\min\{|\det B|:B\) is an invertible \(m\times m\) sub - matrix of \(A\}\).
- When \(X=\mathbb{Z}^n_{\geq0}\), for general values of \(k\), there has been no relevant research yet. But in special cases (for example, \(\text{pos}\{a_1,\dots,a_n\}=\mathbb{R}^m\)), similar lattice - point techniques can be applied.
4. **Semigroup generators**: When \(X = \mathbb{Z}^n_{\geq0}\) and the column vectors of \(A\) generate a pointed cone, by introducing the norm \(\|\cdot\|_{P(B)}\) induced by the symmetric parallelepiped \(P(B)\) and using the parameter \(\mu:=\max_{i\in[n]}\|a_i\|_{P(B)}\), it is proved that \(\text{app}_{\mathbb{Z}^n_{\geq0},k,\|\cdot\|}(A)\) decreases exponentially with \(k\).
### Summary
The main goal of this paper is to study the best possible bounds for approximating \(b\) with sparse solutions \(y\) under different sets \(X\) and norms \(\|\cdot\|\), and prove that as the sparsity \(k\) increases, the approximation quality improves exponentially. This research not only expands the scope of existing sparse recovery problem research, but also provides new insights into understanding the errors in sparse representations.