LAP Dataset

This dataset contains benchmark instances for the Linear Assignment Problem (LAP), generated for evaluating heuristic and exact optimization algorithms. It includes three types of matrices—uniform, structured, and skewed—covering sizes from 10×10 up to 9000×9000, with 30 independently generated instances for each size. Every file is an n×n cost matrix stored as a CSV without headers, named using the format type_size_instance.csv. These datasets were generated using Python random initialization (LAP_Generator.py) for reproducibility and are intended for use in comparing algorithms such as the proposed heuristic, Hungarian, Greedy, Iterated Greedy, Gradient-based heuristic, Simulated Annealing, and Auction algorithm. The dataset supports full reproducibility of the experimental results presented in the associated research study.