MilleMiglia: Bridging the Data Gap in Middle-Mile Logistics Research

Middle-mile logistics—the segment of the supply chain moving bulk goods between distribution centers at regional or continental scale—often flies under the radar. While the first and last miles of logistics (connecting producers to collection points and delivering to consumers, respectively) have been extensively modeled and optimized, the middle mile has received considerably less academic attention, despite being responsible for a sizeable share of overall logistics costs. This neglect is largely due to the absence of accessible, high-quality, and realistic data, as commercial logistics networks are typically treated as proprietary information by companies.

Introducing MilleMiglia: An Open-Source Instance Generator

To address this critical data gap, researchers from Google and academic partners have developed MilleMiglia, an open-source C++ instance generator for the middle-mile logistics segment source. MilleMiglia creates realistic, privacy-preserving synthetic benchmarks for evaluating algorithms and solvers tasked with optimizing complex freight flows across continental networks.

Unlike standard vehicle routing problem (VRP) solvers—which focus on first- and last-mile logistics—MilleMiglia models the unique operational constraints found in the middle mile. These include:

This realism sets MilleMiglia apart, enabling researchers to simulate situations where, for example, a shipment missing its outgoing connection at a distribution center could face significant delays—accurately reflecting real-world logistics complexity.

Data-Driven Benchmarking at Multiple Scales

MilleMiglia produces synthetic data that mimics actual distribution networks by using data-driven statistical distributions, such as gravity models for spatial layout and realistic origin-destination demand patterns. The tool can generate sets that range from small academic test cases ('toy' problems) to industrial-sized, continent-wide networks ["The generator supports creating instances of varying sizes from small academic toy problems to large industrial-scale networks for solver development and machine learning applications."]. This versatility is crucial for both developing new optimization algorithms and training machine learning models at scale.

Importantly, MilleMiglia was designed with privacy in mind. Because it synthesizes rather than exposes real network data, it supports research without risking confidential business information ["MilleMiglia uses data-driven statistical distributions to produce spatially and temporally plausible synthetic data preserving privacy while enabling research in complex routing and synchronization problems."].

Collaboration and Open Science

The MilleMiglia project emerges from a collaboration between Google and academic institutions including UniBrescia and ENPC Paris. Its open-source code and documentation are accessible via GitHub, inviting community engagement and further development ["The project is a collaboration between Google and academic partners, with source code and documentation available on GitHub."].

The creators envision MilleMiglia as the first step toward a standardized middle-mile benchmarking suite—functioning for this domain as CVRPLIB does for classical VRP research. With efforts underway for specialized solvers, this initiative is poised to make middle-mile logistics a fertile ground for academic and industrial innovation.

Implications for AI-Driven Online Businesses

Optimizing the middle mile is essential for online businesses that depend on efficient supply chain operations to deliver value at scale. While MilleMiglia itself is not explicitly aimed at AI-driven business operating systems, its role in enabling realistic, large-scale logistics benchmarking supports the broader trend of integrating advanced optimization and machine learning into operational infrastructure. Solutions like r3to.OS—positioned as an agentic operating system for online businesses—stand to benefit from future research and standardized datasets that MilleMiglia makes possible, especially when orchestrating complex, multi-region delivery operations.

---

Further Reading: