A new paper presented at IPIC 2026 (Errami, Hashemi-Petroodi, Puchinger, Klibi) tackles a problem central to the Physical Internet vision for cities: how do you actually get competing logistics providers and public transport authorities to share infrastructure, not just talk about it?
The model, developed under the EU MODALSHIFT project and tested on a real Madrid case, structures urban freight through three tiers — gateway hubs (city distribution centers), local hubs (intermodal stations), and access hubs (rendez-vous points close to customers). Freight moves from trucks in tier one, gets consolidated into modular π-containers and carried on public transport (trams, metro, buses) in tier two, then reaches customers via city freighters — cargo bikes, vans — in the final tier.
Two contributions stand out. First, the team compares open versus closed-route last-mile delivery and finds open routes (freighters don’t need to return to depot) consistently produce lower-cost solutions, though they’re harder to solve to proven optimality. Second, and more interesting for anyone working on public-private collaboration: the paper introduces fairness constraints borrowed from coalition game theory, ensuring no sub-coalition of logistics providers does better acting alone than within the shared system, while guaranteeing each provider a minimum workload share. Numerical tests with two competing LSPs show real gains from collaboration — up to 24% cost reduction in some instances, averaging 7-8%.
The catch, as the authors are candid about: solving these models at scale is hard. Even small instances (14 customers, a handful of hubs) push commercial solvers to their time limits. Next steps include column generation and branch-cut-price methods, plus adding real-time uncertainty. This is a reminder that the gap between elegant network design and operational deployment in cities remains wide.
Walther Ploos van Amstel.
Also read: How far along is the Physical Internet in city logistics? A Dutch reality check