Bipartite independence number and balanced coloring
Debsoumya Chakraborti
Abstract
Debsoumya Chakraborti
Abstract
In this paper, we establish a couple of results on extremal problems in bipartite graphs. Firstly, we show that every sufficiently large bipartite graph with average degree $\Delta$ and with $n$ vertices on each side has a balanced independent set containing $(1-\epsilon) \frac{\log \Delta}{\Delta} n$ vertices from each side for small $\epsilon > 0$. Secondly, we prove that the vertex set of every sufficiently large balanced bipartite graph with maximum degree at most $\Delta$ can be partitioned into $(1+\epsilon)\frac{\Delta}{\log \Delta}$ balanced independent sets. Both of these results are algorithmic and best possible up to a factor of 2, which might be hard to improve as evidenced by the phenomenon known as `algorithmic barrier' in the literature. The first result improves a recent theorem of Axenovich, Sereni, Snyder, and Weber in a slightly more general setting. The second result improves a theorem of Feige and Kogan about coloring balanced bipartite graphs.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
In this paper, we establish a couple of results on extremal problems in bipartite graphs. Firstly, we show that every sufficiently large bipartite graph with average degree $\Delta$ and with $n$ vertices on each side has a balanced independent set containing $(1-\epsilon) \frac{\log \Delta}{\Delta} n$ vertices from each side for small $\epsilon > 0$. Secondly, we prove that the vertex set of every sufficiently large balanced bipartite graph with maximum degree at most $\Delta$ can be partitioned into $(1+\epsilon)\frac{\Delta}{\log \Delta}$ balanced independent sets. Both of these results are algorithmic and best possible up to a factor of 2, which might be hard to improve as evidenced by the phenomenon known as `algorithmic barrier' in the literature. The first result improves a recent theorem of Axenovich, Sereni, Snyder, and Weber in a slightly more general setting. The second result improves a theorem of Feige and Kogan about coloring balanced bipartite graphs.
Key concepts: Bipartite graph, Combinatorics, Independence number, Mathematics, Independent set, Vertex (graph theory), Degree (music), Discrete mathematics