Vehicle routing optimisation
Shortest route, wrong plan. We plan your whole fleet at once, with all its stops and time windows.
- Free 30-minute call
- No commitment
- You own the code
In short: A vehicle routing optimisation decides which vehicle serves which stops, and in what order, so the whole fleet’s day costs as little as possible. It keeps every load limit, delivery time window and driving-hour rule at once, and knows which depot each vehicle starts from.
The problem
On Wednesday at four, tomorrow’s routes are due. You have the stop list, the fleet and a spreadsheet that’s been patched so often nobody remembers who built it. By seven there’s a plan. It works, too. What does it cost? Was there a cheaper one? You can’t say.
Why the spreadsheet can’t see the whole day
Routing demos love twenty stops on a sunny map. Your Thursday has two hundred, three depots and a customer who won’t take a delivery after nine. Every route depends on the others. If one customer needs an earlier delivery, another depot may be the better one to serve them. A different vehicle leaves from there, and its route changes. In a spreadsheet, you plan one vehicle after the other. So some vans leave half-empty, others run late, and nobody can say what one more customer costs.
What changes
Today, you plan tomorrow’s routes by hand, one van at a time, into the evening. Our model proposes the whole day for every vehicle at once. You check a plan instead of building one, move a stop if needed and send it out. The plan keeps every rule you gave it: loads, time windows, depots, vehicle types and driving hours.
Custom VRP algorithm development: what you get
You see every plan in the numbers you already track: total distance, vehicles used, on-time rate and planning time. Ask what one more customer would cost, and you get an answer. Before anything goes live, we test the model on your own orders. You own the model and the code, so you don’t have to change how you work to fit a packaged tool.
- A model written around your own vehicle routing optimisation rules.
- Results measured in the numbers you already track.
- It slots into how you already work, and you can check every answer it gives.
Driving-time rules belong in the plan
Many route planners treat a driver’s day as one long block. Your drivers don’t live it that way. A day has a cap on driving hours, and after a stretch at the wheel comes a break the driver has to take. Heavy trucks and vans often run under different limits, and your own staff agreements may add more. So the break often comes in the middle of a route, not after the last stop.
Take a truck that leaves at six with stops due by eleven. Ignore the break, and the driver takes it anyway around half past ten and misses the last two time windows. So the break goes into the model as a rule, and the model places it where it costs least. You get a plan your drivers can actually drive, breaks included.
How it works
We write your depots, your customers, your fleet and your rules into one model. You get the routes that cost the least and keep every rule. You still have the last word: you check the result and send it out.
What the vehicle routing problem is
The vehicle routing problem (VRP) asks one question. Which vehicle visits which stops, and in what order? The answer must keep every promise you made and cost as little as possible. George Dantzig and John Ramser first described it in 1959, for petrol deliveries. A map app optimises one trip, while a VRP model plans the day of every vehicle at once.
The variants you run every day
Each everyday rule has its own name in planning. When every truck has a load limit, it’s the capacitated vehicle routing problem (CVRP). A route is only valid if the whole load fits on the truck. Add the delivery windows your customers care about, and you get the VRP with time windows (VRPTW). A delivery that arrives after the customer has stopped taking goods is worthless, however short the route.
With two or more depots, every stop raises a second question: which depot should serve it? That’s the multi-depot VRP (MDVRP). A mixed fleet makes the planning harder still. Vans, trucks and trailers differ in how much they carry, which streets and customers they can reach and what they cost per kilometre. Most businesses have all of this on the same day. And no generic route optimisation algorithm knows when your customers take deliveries. So we solve it in one model, and you get one plan that keeps every rule.
Common questions.
Something else? Ask us directly.
There’s no fixed limit. Up to a few hundred stops, a model can usually prove that a plan is the best possible. For bigger days, it uses heuristics. Those are rules of thumb that find very good plans fast, but without proof. Large operators also split the problem by region and then improve the routes along the region borders. For you, the useful test is simpler: does the new plan beat today’s on your own data?
Usually because the model doesn’t know what the driver knows. Straight-line distances, a missing one-way street or a wrong time per stop all produce routes that look short but are slow to drive. The fix is better data in the model: road distances, measured stop times and rules such as which entrance a truck must use. You should also be able to lock a stop and let the model plan around it.
If you plan alone, with a handful of stops and few hard rules, often yes. Google Maps takes nine stops per route at most, including the destination, and you set the order yourself. A spreadsheet stops working once several vehicles share the work, delivery time windows get tight or the plan changes during the day. If planning eats hours and stops when one person is away, you’ve outgrown it.
It should, and many don’t. The time of the break decides which time windows you still make. So the break belongs inside the route. Driving until the maximum driving time is reached doesn’t always work. An early break can help the driver reach a tight time window later in the day. Vans and heavy trucks often run under different limits, so each vehicle carries its own break rule in the model.
Sometimes it can, and if that’s the case, we’ll tell you. Standard software is built for an average operation, though. Yours runs on its own rules: your contracts, your exceptions, the things you know by heart. A custom model writes those rules in, and that’s where the better plan comes from.
You don’t need either to start. We build the model from the data you’ve already got, even if it lives in three spreadsheets and one person’s head. You get a system you run yourself. We train your team until they no longer need us.
You see a first result with the prototype, before you commit to anything bigger. We start with a free 30-minute call about your operation and the model that would fit it. Then we build a working prototype and run it on your numbers.
Similar pages.
Terms on this page
Every abbreviation explained in plain words, in our article “Metrics and methods”.
- VRPVehicle routing problem
- CVRPCapacitated vehicle routing problem
- VRPTWVehicle routing problem with time windows
- MDVRPMulti-depot vehicle routing problem
- Heuristic and metaheuristic
Tell us what you’re planning.
In a free 30-minute call with a specialist, you find out whether optimisation can help.
- Free, no commitment
- You talk to a specialist directly
- A clear go or no-go