🤖 AI Summary
This study addresses the problem of characterizing unimodularity for graphs: specifically, when a graph’s incidence matrix is totally unimodular, and when its toric ideal is unimodular. Using an integrated approach from algebraic combinatorics and graph theory, we establish an exact structural correspondence—proving that a graph’s incidence matrix is totally unimodular if and only if every pair of odd cycles intersects (i.e., has nonempty intersection). Consequently, we derive necessary and sufficient graph-theoretic conditions for the unimodularity of its toric ideal. This result constitutes the first complete reduction of unimodularity to the topological intersection pattern of odd cycles, yielding an exact equivalence between an algebraic property and a combinatorial structure. The work fully classifies unimodular graphs, and provides foundational criteria and constructive tools for verifying total unimodularity of constraint matrices in integer programming and for algebraic representations of graphs.
📝 Abstract
We give a necessary and sufficient graph-theoretic characterization of toric ideals of graphs that are unimodular. As a direct consequence, we provide the structure of unimodular graphs by proving that the incidence matrix of a graph $G$ is unimodular if and only if any two odd cycles of $G$ intersect.