The Yang-Baxter equation is a fundamental relation that appears in several areas of physics and mathematics, including quantum mechanics, statistical mechanics, and the study of braids and knots. In its matrix version, the equation asks you to find all matrices X that satisfy a specific relationship with a given fixed matrix A, written as AXA = XAX. Solving this equation is generally quite difficult, and researchers have been systematically working through special cases by choosing particular types of matrices for A and cataloging all valid solutions X.
This paper focuses on the case where A is a permutation matrix, meaning it is a square matrix filled with zeros and ones where each row and each column contains exactly one entry equal to one. Such matrices represent rearrangements of coordinates. The authors look specifically for solutions X that are doubly stochastic, meaning every row and every column of X sums to one and all entries are non-negative. Doubly stochastic matrices are a well-studied class that includes permutation matrices as special cases but also contains many other matrices with fractional entries, and they arise naturally in probability and optimization.
The paper fully characterizes all doubly stochastic solutions to the equation for any permutation matrix A. The authors break the solutions into two groups: those X that are themselves permutation matrices, and those that are genuinely fractional doubly stochastic matrices (called proper doubly stochastic solutions). They provide explicit descriptions of both types and illustrate their findings with concrete examples. The result gives a complete picture of how the symmetry structure of a permutation matrix A shapes and constrains all the doubly stochastic matrices that can satisfy the Yang-Baxter-like relation with it.