Routes built against the day you actually have
Fixed weekly routes drift out of date the moment volumes, windows, or vehicle availability move. Solving daily against the real order book recovers the distance those assumptions cost.
Fuller trucks, fewer kilometres.
Routing, load building, and network design solved together rather than in sequence. The cost of a delivery is largely fixed by decisions made before the vehicle is loaded.
Three problems worth modelling
Trucks that leave full, legally
Load building is a packing problem under axle weights, stacking rules, and delivery sequence. Treated as one, the same fleet carries measurably more per trip.
A network sized for where demand is now
Hub locations, lane structure, and fleet mix are usually inherited rather than chosen. Modelling them exposes what the current shape costs against the one the demand justifies.
What we build
- Vehicle routing with time windows, capacity, and driver-hours constraints
- Load and container planning under weight distribution and stacking rules
- Network and hub-location design across lanes, depots, and cross-docks
- Fleet sizing and mix, including own-versus-hired trade-offs
- Carrier allocation and freight procurement across rate structures
- Slotting and pick-path optimisation inside the warehouse
Methods: VRP / VRPTW, MILP, column generation, bin packing, network flows
Outcomes
- Fewer kilometres: the same drops served on shorter, denser routes
- Higher fill rate: loads built as a packing problem, not by eye
- Defensible network spend: what a depot or lane is worth, quantified before it is signed