Multi-Stop Route Planner Software for Last-Mile Delivery: Route Optimization Explained

A multi-stop route planner, the core of last-mile route optimization software, decides which van takes which stops, and in what order, so the day's total time or distance is as small as possible and every time window is met. A map app does point-to-point routing, the fastest way from here to there. Multi-stop planning starts from sixty stops, several vans, time windows on many of the stops, van capacities, driver shifts and a depot. The number of possible plans is astronomical: a single van with sixty stops has more possible orders than there are atoms in the universe. The software doesn't check them all; it searches cleverly and finds a very good plan in seconds. Last-mile delivery is the case route optimization was built for.

A last-mile plan has to respect all of these constraints at once.

In the morning the software builds routes from the day's manifest and sends each driver's app the stops in order, with navigation. That saves the planner an hour and produces a better plan. Most of the day's value, though, comes from the re-plan: a cancelled stop, a new pickup, a van running behind at 11am. Good last-mile optimization re-sequences a driver's remaining stops in seconds and can move stops between vans, while the driver's app updates and the customer's arrival estimate is recalculated. When you talk to vendors, ask them to demonstrate the re-plan instead of the morning plan.

Optimization is only as good as its inputs, and the best inputs are your fleet's own telematics history: how long this driver spends at this kind of stop, how long this street takes at this hour, where the van was when the stop was marked delivered. A route optimization tool that reads your telematics learns these over weeks and stops getting the same street wrong every morning. Live vehicle location is also what tells the tool a van is behind before the driver calls in.

Measure stops per route per hour, distance per stop, on-time rate against windows and planner time. The gains to look for when moving from manual planning to optimization are more stops per route, fewer kilometres per stop and a planner freed up for exceptions. Hold a vendor to those numbers in a trial by running the same day's manifest both ways and comparing the results.

A fair test of a last-mile optimizer takes one week and your fleet's own data. Take five days of real manifests and plan each day two ways, the planner's way and the software's, then run the software's plan on two vans while the rest run as usual. At the end, compare four numbers: stops per route, kilometres per stop, on-time rate against customer windows and the planner's hours. On day three, test the re-plan by adding six pickups at 10 a.m. and cancelling four stops, and time how long the driver's app takes to update. Ask to see the plan for the same day a week later, after the tool has learned your real stop durations, because its first plan is the worst one it will make. High Point GPS runs this trial with routing partners integrated with the High Point GPS platform, for fleets across Canada and the United States.