How Bike-Share Rebalancing Trucks Decide Where to Move Bikes

It's 7:38 on a Wednesday and the dock outside your building has nineteen empty slots and zero bikes. Four blocks downhill there's a station so overfull that people are leaning bikes against the rack. Nobody's incompetent. This is the system in its normal state, and the trucks driving around all day exist to fight it.

It's mostly gravity and commuting

Two forces empty and fill stations, and they both point the same direction every weekday morning.

The first is directional commuting. Everybody rides from residential areas toward job centers between roughly 7 and 9am, then reverses in the evening. Demand isn't symmetric within any given hour, so origin stations drain and destination stations saturate. Simple, relentless.

The second is literal gravity. In any city with hills, bikes accumulate at the bottom. Riders will happily coast down and take transit or walk back up, so the uphill stations bleed inventory all day. Systems in hilly cities spend a disproportionate share of their rebalancing budget hauling bikes back uphill, and e-bikes reduce this effect without eliminating it.

Add a stadium event dumping a hundred bikes at one station in twenty minutes, or a rainstorm that strands inventory wherever it fell, and you have a distribution that never sits still.

What the dispatcher is looking at

A console shows every station with its bike count, dock capacity, and fill percentage. Stations flag themselves when they cross thresholds — near-empty, near-full — and most systems overlay a short-horizon forecast of where each station is heading, built from historical patterns for that day and time, adjusted for weather.

The dispatcher also sees what the model doesn't: which vans are staffed today, where each one is, how much shift time is left, and which streets are closed for a parade. The binding constraint is usually van-hours remaining before shift end.

The math that picks the route

Formally this is a pickup-and-delivery vehicle routing problem with inventory targets. In plain terms: a van with limited capacity — commonly somewhere in the range of fifteen to twenty-five bikes depending on the vehicle and rack design — has to visit a sequence of stations, taking bikes from surplus ones and dropping them at deficit ones, minimizing driving time while getting inventory close to target levels before the demand peak hits.

Two details in that sentence do a lot of work.

Before the peak, not during it. A van arriving at a commuter origin station at 8:15 is too late; the demand already happened and the riders gave up. That's why the real work happens overnight and pre-dawn, when a truck can reset dozens of stations without fighting traffic.

Targets aren't "full." Target inventory varies by time of day, and for a commuter destination it might be near zero in the morning — because it's about to receive a flood of arriving bikes and needs empty docks. A station full of bikes at 8:45 in a job center is a failure, not a success.

Here's the part that surprises people: operators deliberately let some stations sit empty. If a station generates two rides a day and sits twenty minutes off any efficient route, servicing it costs more van time than it returns in trips. The objective function is unmet demand across the network, not fairness across stations — which is exactly why cities increasingly write service-level requirements into contracts for lower-income neighborhoods, because pure efficiency optimization will quietly abandon them.

Every tool in the rebalancing kit

Method	How it works	Best for	Limitation
Overnight van or box truck	Large batch resets while the network is idle	Getting the whole system to morning targets	Labor cost; can't respond to anything that happens after
Daytime van circuits	Reactive runs triggered by empty and full alerts	Correcting mid-day imbalance and event surges	Slow in traffic; parking the van is its own problem
Bike trailer or cargo bike	Small hauls, a handful of bikes at a time	Dense cores where a van can't park or maneuver	Low capacity per trip
Station valet	Staff on site at a chronically saturated station, receiving and clearing bikes live	Major transit hubs and event venues at peak	Expensive; only worth it at extreme-volume locations
Rider incentives	Credits, free minutes, or points for ending a ride at a station that needs bikes	Broad, cheap nudging at scale	Riders only detour so far; unreliable for a specific station by a specific time
Virtual or flexible stations	Temporary designated drop areas to absorb overflow	Events and construction closures	Needs enforcement or it becomes clutter
Dynamic pricing	Surcharges or discounts by station to shape where trips end	Systems with app-based fare control	Equity concerns; can read as punishing riders for the operator's problem

Why paying riders beats driving a van

A van rebalancing a bike involves fuel, a driver's time, traffic, and a parking maneuver. A rider rebalancing the same bike involves a small credit and a two-block detour they were half willing to make anyway. On pure cost per bike moved, incentives usually win by a lot.

They just can't be steered precisely. Incentives shift a distribution; they don't guarantee that this station has eight bikes by 7:30. So most mature systems run both: incentives as continuous background pressure, vans for the specific commitments that have to be met.

E-bikes changed the problem entirely

The moment a fleet goes electric, rebalancing stops being the dominant constraint and charging takes over.

Now a bike can be in exactly the right place and still be useless, because it's at eight percent. Staff are either swapping packs in the field or hauling bikes to a depot, and in systems with charging docks the bike needs a powered slot, not just any slot. The fleet has a second inventory dimension — location and charge — and a van route has to serve both. Field swaps win on efficiency because the bike never leaves service, which is why so many operators moved to swappable packs.

Free-floating fleets face the same problem without the docks to organize it, which is how you end up with clusters of dead scooters and bikes in low-demand corners waiting for someone to come find them.

How to stop getting burned

Theory aside, here's what works as a rider.

• Check the app for dock status before you walk to the station, not after. Most apps show live bike and open-slot counts, and the walk to a second station is much shorter from your door than from an empty rack.
• If you're riding into a job center in the morning, check open docks at your destination too. Getting there is only half the trip.
• Learn the uphill station near you and stop trusting it at peak. Downhill stations run full; uphill ones run empty.
• Turn on incentive notifications if your system offers them. Free credit for the detour you were already making is the best deal in bikeshare.
• If your station is chronically dead at the hour you need it, report it with the specific station, day, and time. That's the input operators use to adjust targets, and vague complaints get nothing.

And if a station is empty three mornings running, treat it as permanent until proven otherwise. Build the backup into your commute instead of discovering it in the rain.