A Benchmark Corpus of Realistic Logistic Transport Networks with Nonlinear Edge Costs and Transshipment Capacities
收藏资源简介:
This dataset is an open corpus of benchmark instances for logistic transport network flow problems with nonlinear (vehicle-induced) edge costs and hard storage transshipment capacities. It accompanies the paper "A Benchmark Generator of Realistic Logistic Transport Networks with Nonlinear Edge Costs and Transshipment Capacities." The instances are produced by a parameterised generator calibrated on real road graphs from OpenStreetMap. The generation pipeline combines Zipf-distributed city sizes, a spatial-network budget topology model, centrality-driven storage capacities, and a doubly constrained gravity demand model, yielding networks whose structural statistics match those of real regional transport systems. Contents (50 instances, 10–150 storages): Synthetic/ — 35 fully synthetic instances (<N>_nodes_<k>, where N is the storage count and k the replicate index), spanning 10 to 150 storages with multiple replicates per size. Real_Graph/ — 15 real-geometry instances built on actual regional road graphs: Belgium, California, Denmark, European Russia, Georgia, Germany, Hungary, Illinois, Italy, Netherlands, Ohio, Spain, Sweden, Texas, and the United Kingdom. Per-instance files: pos_pop.csv — node coordinates and population: x, y, population distance_matrix.csv — pairwise edge prices: src, dst, price reqs.csv — transport demands: src, dst, volume offices.csv — storage/transshipment parameters: office_id, transfer_price, transfer_max graph.graphml — the network in GraphML format data.txt — the complete generation parameters for the instance (including the random_seed), making each instance exactly reproducible Provenance and credit: This benchmark corpus was designed and created by Varvara Kovaleva as part of her master's thesis at HSE University (Laboratory of Algorithms and Technologies for Network Analysis, Nizhny Novgorod), under the supervision of Alexander Ponomarenko, who proposed the underlying optimal-logistics problem. The original generator and dataset repository is maintained by Varvara Kovaleva: https://github.com/Varya-K/Optimal-Logistic-Net . A derived version is maintained at https://github.com/aponom84/Optimal-Logistic-Net . Please cite: V. Kovaleva and A. Ponomarenko, A Benchmark Generator of Realistic Logistic Transport Networks with Nonlinear Edge Costs and Transshipment Capacities, HSE University, 2026.



