"HERO Benchmark Dataset: Experimental Results for Fair Vehicle Routing with HNSW-Accelerated ALNS"

"This dataset contains comprehensive experimental results for fairness-aware Vehicle Routing Problems (VRP) with Time Windows, comparing five optimization methods across multiple benchmark instances. The dataset includes 1,500 successful experimental runs (300 per method) with 10 random seeds for statistical significance.Methods Compared:- ALNS: Classical Adaptive Large Neighborhood Search (baseline)- ALNS-HNSW: ALNS with HNSW-accelerated candidate generation- HERO: Hierarchical Equitable Routing Optimization (HNSW-accelerated ALNS with fairness-aware features)- OR-Tools: Google OR-Tools exact solver (time-limited)- PyVRP: PyVRP exact solverBenchmark Instances:- Solomon VRPTW instances (100 customers): C101-C109, C201-C208, R101-R112, R201-R211, RC101-RC108, RC201-RC208- Homberger extended instances (200-1000 customers): Multiple instance types across different sizes- CVRP X instances (101-1001 customers): Various instance sizes- Euro-NeurIPS 2022 real-world instances (200-880 customers)Metrics Included:- Solution cost (total routing cost)- Number of routes- Coefficient of Variation (CV) - driver workload fairness metric- Jain's Index - route cost distribution fairness metric (computed as J = 1\/(1+CV\u00b2))- Runtime (seconds)- Instance metadata (name, type, number of customers)Data Format:- JSON files: Detailed results with full metadata per run- CSV files: Tabular format for easy analysis (instance_name, instance_type, n_customers, method, seed, cost, n_routes, cv, time_seconds)- Summary files: Aggregated statistics by methodExperimental Setup:- 1,500 iterations per run for ALNS-based methods- 10 random seeds (42, 123, 456, 789, 1011, 1213, 1415, 1617, 1819, 2021) for statistical significance- All methods evaluated on identical instance sets and data splits- Fairness metrics computed using coefficient of variation (CV) of route costs and Jain's IndexThis dataset enables researchers to:- Reproduce and verify the experimental results- Compare fairness-aware optimization methods- Analyze the trade-off between cost efficiency and fairness- Evaluate HNSW acceleration effectiveness for VRP- Conduct further research on equitable vehicle routing "