Regular simple path enumeration is a fundamental problem in heterogeneous information networks (HIN), with important applications in finance, security, and neuroscience, and textile structure analysis. In textile systems, different yarn types (e.g., warp, weft, and conductive fibers) can be naturally modeled as heterogeneous entities, while their interactions (e.g., over, under, and contact relations) form heterogeneous links, making HINs a suitable representation for complex fabric structures and conductive textile pathways. In this work, we study the regular simple path enumeration problem. A straightforward approach is to use a search-based algorithm to enumerate all paths with regular constraints and verify path simplicity. Large, densely connected graphs present unique challenges. In this paper, we propose an efficient graph relabeling algorithm to reduce computation. We also develop new techniques to improve performance. These techniques enhance reachability detection, optimize candidate vertex search, and minimize redundant path computation. Extensive experiments demonstrate that our method can compute the exact result for different kinds of regular path queries efficiently.

heterogeneous information networks, path enumeration, DFS algorithm, structural analysis