Sharing 20 Paris station visits between four crews
Four crews share 20 Paris station visits, five each. We changed which crew gets each station to reduce the distance from its base.
Distribute visits to crews in turn.
Assign visits by distance from each base.
Dots are stations. Houses are crew bases.
The same five visits per crew, closer to base
Total base-to-station distance falls by 12.6%. The crews and inspection task are assumed; these are assignments, not daily routes or measured working hours.
Data, assumptions & full results
Data behind this study
Paris Data · Vélib station locations ↗Real station locations, retrieved on 18 October 2026. The inspection task and team bases are assumptions for this study.
Smovengo / Syndicat Autolib’ Vélib’ Métropole · ODbL 1.0
Assumptions & verification
The first four published stations in station-code order become assumed bases; the next 20 become visits. The reference assigns them to crews A, B, C and D in turn. The model keeps exactly five visits per crew and minimizes the sum of one-way, straight-line base-to-station distances. It does not sequence visits. These hypothetical crews do not represent Vélib staffing.
Every displayed plan was rescored and independently checked. Three-second search budget; no proof of optimality. The downloads include inputs, assignments and verification records.
Model: Resource Allocation