Google Research open-sources MilleMiglia, a C++ generator for realistic middle-mile logistics benchmarks

Google Research published a blog post and paper announcing MilleMiglia, an open-source C++ tool that generates synthetic "middle-mile" logistics network benchmarks. It creates distribution centre placements, shipment demand patterns and vehicle rotation schedules using statistical models (gravity/spatial clustering), framed as a multi-commodity flow problem. Code is on GitHub alongside the paper, authored by Aymane Lotfi and Thibaut Cuvelier.

This is a research tool, not a product or service. It gives academics and operations-research teams a way to test logistics optimisation algorithms without needing proprietary network data from real logistics companies, which has historically been the bottleneck for this niche of research. The people affected are a narrow group: OR academics, logistics-optimisation vendors and possibly Google's own teams (OR-Tools, Route Optimization). There is no evidence yet that any company has changed how it plans logistics networks because of this.

The framing of this as a "foundational building block" for "more robust, efficient global supply chains" is unearned marketing language. A synthetic benchmark generator does not itself improve supply chains; it only helps if researchers adopt it, publish comparable results on it, and those results eventually influence real routing software. That chain of "if" is doing all the work, and none of it has happened yet.

1. 01GitHub activity on the MilleMiglia repo (stars, forks, issues) over the next 3-6 months as a proxy for research community uptake
2. 02Citations of the accompanying paper or use of MilleMiglia benchmarks in subsequent OR conference papers (e.g. INFORMS, EURO) within a year
3. 03Whether Google's own tools (OR-Tools, Maps Platform Route Optimization) publish benchmark results using MilleMiglia, which would signal internal as well as external adoption