Multi-Depot Route Optimization: How to Plan Deliveries From More Than One Base
Multi-depot route optimization assigns each delivery to a depot and a vehicle, and orders each vehicle's stops, so that total cost across all depots is as low as possible. Simple setups split stops by fixed depot catchment areas and then optimize each route. Full multi-depot solvers decide depot assignment and routes together, which mainly pays off for stops near the boundary between catchments.
This article explains the problem, a practical method for small multi-site fleets, and when a full multi-depot solver is worth it. For the wider context, see route optimization software for delivery fleets.
What is a multi-depot routing problem?
In a single-depot problem, every vehicle starts and ends at one base. In a multi-depot vehicle routing problem (MDVRP), vehicles are based at several sites: warehouses, stores, cross-docks or drivers' homes. The solver must decide:
- Which depot serves each stop.
- Which vehicle at that depot serves it.
- In what order each vehicle visits its stops.
It is a standard variant of the vehicle routing problem studied in operations research, and open-source solvers such as Google OR-Tools can model vehicles with different start and end locations.
Three ways to plan multi-depot deliveries
| Approach | How it works | Pros | Cons |
|---|---|---|---|
| Fixed catchments | Each depot owns an area (postcodes or suburbs); stops go to their area's depot, then each route is optimized | Simple, predictable, drivers learn their area | Boundary stops may go to the "wrong" depot |
| Nearest depot | Each stop goes to the depot with the shortest driving time, then routes are optimized | Easy to automate | Ignores depot capacity and driver availability |
| Full multi-depot optimization | Solver assigns stops to depots and vehicles and orders routes in one step | Lowest total cost, balances capacity | Needs a fleet platform; plans change day to day |
For a small business with two or three sites, fixed catchments plus per-route optimization usually captures most of the benefit.
Worked example: two depots, one boundary
A florist has a shop in the north (depot A) and one in the south (depot B), with one driver each. Today: 18 deliveries in the north, 15 in the south, and 5 in a suburb halfway between.
- Fixed catchments put the 5 middle stops in A's area: A's driver has 23 stops, B's driver 15. A finishes late.
- Nearest depot by driving time splits the middle stops 3 to A and 2 to B: 21 and 17 stops.
- Full multi-depot optimization might give all 5 to B, because B's route already passes nearby: 18 and 20 stops, with less total driving and both drivers finishing at a similar time.
The illustrative point: the difference between approaches is almost entirely about the boundary stops.
Where full multi-depot optimization helps
- Many stops near catchment boundaries.
- Depots with very different workloads from day to day.
- Vehicle or depot capacity limits (a small store can only load so much).
- Drivers who can start at one site and finish at another.
- Mid-route reloads: vehicles returning to the nearest depot to reload. Some platforms offer this; OptimoRoute, for example, lists "Dynamic Mid-Route Depots" on its Custom plan. See OptimoRoute's mid-route depots.
If your depots are far apart with few shared areas, fixed catchments are usually fine.
Start and end points
Multi-depot planning is also about where routes end:
- Closed routes: start and end at the same depot.
- Depot to depot: start at depot A, finish at depot B (for example to restock B).
- Open routes: start at a depot, finish at the last stop (common for drivers who go home).
Each choice changes the optimal order, so set the real end point before optimizing.
A practical method for small multi-site fleets
- Draw catchments by suburb or postcode, using driving time from each depot rather than straight-line distance.
- Flag boundary suburbs and review them daily: assign them to whichever depot has capacity and a route passing nearby.
- Optimize each driver's route with the right start (their depot) and end (depot, other depot, home or last stop). See optimizing each multi-stop route.
- Check route totals against shift length and move a few boundary stops if one driver is overloaded.
- Review monthly: if boundary decisions take more than a few minutes a day, a full multi-depot platform may pay off; see evaluating platforms for logistics companies.
For long distances between regional depots, see also regional Australian runs.
Using MiliRoute with more than one depot
MiliRoute is an AI route optimization app for couriers, delivery drivers and small delivery fleets. It optimizes one route per driver, so it fits the fixed catchment method above:
- Set each route's start to the driver's depot address (or their current GPS location).
- Set the end to the same depot, another depot, home, or no end location (finish at the last stop).
- Optimize: the stop order is calculated on estimated driving times from Google's road-routing data, with avoid-tolls if you want.
- Move boundary stops between drivers by adding them to the other driver's route, or share a route as a link.
MiliRoute does not assign stops to depots or vehicles automatically; full multi-depot assignment needs a fleet platform. Free for routes of up to 10 stops; Premium A$15/month.
Conclusion
Multi-depot routing adds one decision on top of ordinary route optimization: which depot serves each stop. For most small multi-site businesses, sensible catchments plus a well-optimized route per driver is enough; the gains from full multi-depot solvers come from boundary stops, capacity balancing and reloads. Measure how much time boundary decisions take you before buying a bigger platform.
Optimize each depot's routes: create a free MiliRoute account, set the depot as the start, and optimize.
Frequently asked questions
What is multi-depot route optimization?
It is planning delivery routes when vehicles start from more than one base: assigning each stop to a depot and vehicle, and ordering each route, to minimise total cost.
What is the MDVRP?
The multi-depot vehicle routing problem: a version of the vehicle routing problem in which vehicles are based at several depots.
How do I split deliveries between two warehouses?
Start with catchment areas based on driving time from each warehouse, then review stops near the boundary each day and assign them to the depot whose routes pass nearby or have spare capacity.
Can a driver start at one depot and finish at another?
In many route planners, yes. In MiliRoute you can set the start to one depot and the end to another address.
When is a full multi-depot solver worth it?
When many stops sit near catchment boundaries, workloads swing between depots, depots have capacity limits, or vehicles reload mid-route.
Does the end point matter in multi-depot routing?
Yes. A route that must return to its depot, end at a different depot or finish at the last stop will be ordered differently.