Aida Abiad,Alexander L. Gavrilyuk,Antonina P. Khramova,Ilia Ponomarenko
Abstract:We derive a linear programming bound on the maximum cardinality of error-correcting codes in the sum-rank metric. Based on computational experiments on relatively small instances, we observe that the obtained bounds outperform all previously known bounds.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to develop a method to estimate the maximum cardinality of sum - rank metric codes with a given minimum distance using Delsarte's linear programming (LP) method. Specifically, the authors focus on how to construct a suitable association scheme in the sum - rank metric space and obtain an upper bound by solving the corresponding linear programming problem. Through computational experiments, they observe that the new LP upper bound is better than all previously known upper bounds.
### Background and Motivation
1. **Sum - rank metric codes**: Sum - rank metric codes are a special class of matrix tuple sets, where each matrix is defined over a finite field \( \mathbb{F}_q \). The distance between two such tuples is the sum of the ranks of the differences of their corresponding elements. This metric is a generalization of the rank metric and the Hamming metric.
2. **Central problem in coding theory**: A core problem in coding theory is to determine the maximum cardinality of a code with a given minimum distance. This problem has also been studied in the context of sum - rank metrics, but current methods have not fully solved the problem.
3. **Limitations of existing methods**: Although Delsarte's linear programming method has been successfully applied in the Hamming metric and some other metrics, in the sum - rank metric, since the sum - rank metric graph is usually not a distance - regular graph, this method has not been fully developed.
### Main contributions of the paper
1. **Constructing an association scheme**: The authors propose a method to construct a suitable association scheme by regarding the sum - rank metric graph as the Cartesian product of smaller rank metric graphs. Specifically, they utilize the structure of the sum - rank metric graph, represent it as the direct product of smaller rank metric graphs, and consider the direct product of the association schemes of these graphs.
2. **Linear programming upper bound**: Based on the constructed association scheme, the authors derive a linear programming problem, and by solving this problem, an upper bound of the maximum cardinality of sum - rank metric codes can be obtained.
3. **Computational experiments**: Through computational experiments, the authors find that the new LP upper bound is better than all previously known upper bounds.
### Method overview
1. **Structure of the sum - rank metric graph**: The sum - rank metric graph can be represented as the Cartesian product of multiple smaller rank metric graphs. Each smaller rank metric graph corresponds to a bilinear form graph.
2. **Construction of the association scheme**: By considering the direct product of the association schemes of these bilinear form graphs, construct an association scheme suitable for the sum - rank metric graph.
3. **Linear programming problem**: Using the constructed association scheme, define a linear programming problem with the objective of maximizing the size of the code while satisfying the linear constraints derived from the properties of the association scheme.
4. **Computational results**: Verify the effectiveness of the new method through computational experiments, and the results show that the new method is better than other existing methods.
### Conclusion
This paper successfully obtains a new upper bound of the maximum cardinality of sum - rank metric codes by constructing an association scheme of the sum - rank metric graph and using Delsarte's linear programming method. This result not only promotes the research on sum - rank metric codes, but also provides new tools and methods for further research in related fields.