SVRPBench: A Realistic Benchmark for Stochastic Vehicle Routing Problem

2025-05-28

AI Summary

This work addresses the limited generalization of routing solvers under high-fidelity stochastic dynamics prevalent in real-world logistics—such as time-varying traffic congestion, log-normally distributed delays, probabilistic accidents, and realistic time windows. To this end, we introduce the first city-scale stochastic vehicle routing benchmark, comprising over 500 instances with thousands of customers, systematically integrating heterogeneous uncertainty sources and enabling out-of-distribution generalization evaluation. We propose a reproducible pipeline for generating multi-depot/multi-vehicle stochastic scenarios, combining empirical data-driven modeling, time-dependent graph construction, and constraint-aware instance generation. A standardized open-source evaluation suite and unified interface are provided. Experiments reveal that leading reinforcement learning solvers (e.g., POMO, Attention Model) suffer over 20% performance degradation under distributional shift, whereas classical methods exhibit greater robustness. All datasets, code, and tools are publicly released.

Technology Category

• Planning, Routing, and Scheduling: Scheduling under Uncertainty
• Reasoning under Uncertainty: Stochastic Optimization
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

Robust routing under uncertainty is central to real-world logistics, yet most benchmarks assume static, idealized settings. We present SVRPBench, the first open benchmark to capture high-fidelity stochastic dynamics in vehicle routing at urban scale. Spanning more than 500 instances with up to 1000 customers, it simulates realistic delivery conditions: time-dependent congestion, log-normal delays, probabilistic accidents, and empirically grounded time windows for residential and commercial clients. Our pipeline generates diverse, constraint-rich scenarios, including multi-depot and multi-vehicle setups. Benchmarking reveals that state-of-the-art RL solvers like POMO and AM degrade by over 20% under distributional shift, while classical and metaheuristic methods remain robust. To enable reproducible research, we release the dataset and evaluation suite. SVRPBench challenges the community to design solvers that generalize beyond synthetic assumptions and adapt to real-world uncertainty.

Problem

• Benchmarking vehicle routing under realistic stochastic dynamics
• Evaluating solver robustness against distributional shift in logistics
• Providing diverse urban-scale scenarios with empirical constraints

Innovation

• Open benchmark for stochastic vehicle routing
• Simulates realistic urban delivery conditions
• Evaluates solver robustness under uncertainty