Replication Package for "Recourse-Aware Online Large-Neighborhood Search for the Dynamic Maximum-Weight Independent Set Problem"
Description
This repository provides the complete replication package for the paper "Recourse-Aware Online Large-Neighborhood Search for the Dynamic Maximum-Weight Independent Set Problem" by Jing Li et al. It contains the Python source code for the proposed RA-OLNS framework and comparison baselines (Greedy, periodic recomputation, and time-limited global CP-SAT reoptimization), dynamic graph generator scripts for synthetic (ER, BA, RGG) and real-world SNAP networks under dynamic edge update streams, automated experimental runners across all testing phases, and analysis scripts to fully reproduce all figures, LaTeX tables, and trajectory evaluations reported in the manuscript. Version 2 updates figure-layout settings in the plotting script to prevent label and legend overlap. Algorithms, experimental data, numerical results, and statistical analyses are unchanged.