Argentina's Superliga scheduled its season with cluster patterns instead of place names

Scheduling Argentina's First Division professional soccer league, the Superliga, was a political exercise as much as a logistical one: 20+ clubs across a vast country, broadcasters with slot rights, government requirements, and international windows. Past calendars were built manually and satisfied almost nobody.

The team's integer programming model made geographic fairness structural rather than incidental. Instead of starting from place names, it created and assigned cluster patterns — groups of teams that travel well together — inside a decomposition that then assigned every match to a date and time slot, considering broadcaster conditions, government demands and the international tournament calendar simultaneously.

The model produced clear improvements on multiple criteria over the previous approach and was actually implemented, scheduling the Superliga's 2018–2019 and 2019–2020 seasons.

Cluster teams by geography inside the model, not outside it

Cluster patterns bundle traveling teams into coherent groups, so travel fairness is baked into the optimization rather than patched on after the fact.

Modeling date-and-slot assignment jointly with fixture generation means broadcast and government constraints shape the calendar from the start.

Getting implemented for two consecutive seasons — in a sport where calendars are fought over publicly — is the strongest proof the method worked.

When 'fairness' means geography, encode the geography into the solving structure itself — place-name heuristics fail where cluster patterns let the optimizer honor every constraint at once.

The work was a finalist for the 2020 Daniel H. Wagner Prize and published in the INFORMS Journal on Applied Analytics in 2021. Superliga used the model's schedules for the 2018–2019 and 2019–2020 seasons.