LP Mining with LP2Graph is a novel approach presented by researchers Jörn Maurischat, Nikola Bešinović, and Michael Färber, aimed at improving railway rescheduling. This method was discussed in a paper submitted on July 13, 2026, to arXiv, addressing the challenges in optimizing railway operations.
Understanding LP Mining and Its Application
The field of railway rescheduling heavily relies on Mixed-Integer Linear Programming (MILP). However, the existing modeling knowledge is fragmented across numerous papers, often presented in incompatible notations. The authors propose LP Mining with LP2Graph, a method that systematically extracts and organizes these formulations into a cohesive dataset.
LP2Graph operates by parsing each formulation into a canonical model, creating a structured representation as a typed variable-equation graph. This allows for the deterministic extraction of information, facilitating a clear taxonomy of models that can be applied to railway rescheduling.
Methodology and Validation of LP2Graph
The core methodology involves clustering the extracted data based on variables, constraints, and objectives. This bottom-up approach not only organizes the information by application domain but also by the solution method used. A self-updating classifier labels these groups, enhancing the reproducibility of the dataset.
