Pickup and delivery optimisation
Every pickup has a delivery, and each has its own time window. We plan both ends together, on the same vehicle, in the right order.
- Free 30-minute call
- No commitment
- You own the code
In short: A pickup and delivery optimisation plans routes for loads that one vehicle collects at one place and drops at another. It keeps each pickup before its delivery, meets the time windows at both ends and never overloads the vehicle.
The problem
At eight in the morning, a courier van collects blood samples at a clinic outside town. They have to reach the lab before noon. On the way, the same van picks up two parcels and drops a third. At ten a new pickup comes in. Which van can still take it without missing anybody’s deadline?
Why ordinary routing gets pickups wrong
A parcel you collect has to reach its recipient by a set time. One more rule applies, and ordinary routing doesn’t know it: the item has to be picked up before it can be dropped off. The pickup and the delivery also each have their own time window. Plain routing treats every stop as standing alone, so it misses all of this.
Signs your pickup routing is held together by hand
You keep pickups and deliveries in two lists and pair them up by hand. Routes that look short on screen no longer work after the first hour. Some vans run full in one direction and nearly empty in the other. Miss one pickup window, and the delivery that depends on it is late too. You only find out when the lab calls.
What changes
Today, you pair pickups with deliveries by hand and hope the windows line up. Our model treats each pickup and its delivery as one pair. It puts both on the same van, in the right order, and checks the load after every stop. A new pickup at ten goes to the van that can still make both windows.
Pickup and delivery optimisation: what you get
We build on vehicle routing optimisation and on replanning during the day, and apply both to your pickup and delivery orders. Three rules always apply: pickup comes before delivery, the load fits after every stop, and both time windows are kept. In one step, we decide which vehicle takes which pair and which shipments travel together. You judge the result on vehicles needed, total distance, on-time collection and on-time delivery. For medical and other deadline-critical work, we test on your own orders and peak days first.
- A model written around your own pickup and delivery 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.
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 PDPTW is
In optimisation, this task is called the pickup and delivery problem with time windows (PDPTW). Stops come in pairs: one pickup and its delivery. A PDPTW algorithm has to keep three rules at once. Pickup comes before delivery. The load must fit after every stop. And both the pickup and the delivery have a time window.
Precedence: pickup before delivery
The defining rule is precedence, which means one stop must come before another. Each delivery is the partner of one specific pickup. You can’t deliver before you collect, so both stops go on the same vehicle, in that order. A route can be short and sensibly ordered and still be impossible, because a pickup and its delivery won’t fit inside their windows.
Capacity that changes along the route
On a normal delivery round, the load only goes down. In PDPTW it moves both ways, so the load has to fit after every stop, not just at the start. Take a van with room for ten crates. It leaves with six, picks up five at the first stop and drops two at the second. Check only the start and the end, and the route looks fine. At stop one the van needs eleven places, and the plan doesn’t work.
Time windows at both ends
The pickup has its own window, just like the delivery. You can only collect the sample once the clinic has opened. You have to collect the furniture before the customer leaves for work. Miss the pickup window, and you can’t reach the delivery window any more. Both windows belong to the same trip. If the pickup has to happen later, less time is left for the delivery. That’s why we solve pickup order, delivery order and all the windows together.
When the pickup point is a choice
Textbook pairs are fixed: this pickup, that delivery. In practice they often aren’t. A spare part may be stocked in two warehouses, and either can supply the job. The usual shortcut picks the nearest warehouse first and routes afterwards. That shortcut misses the case where a van already drives right past the warehouse that is farther away.
A better model treats the pickup location as a decision. It lists the possible warehouses and chooses exactly one. For that warehouse, the same rules apply: pickup first, then delivery. Standard pickup-and-delivery models know only one fixed pickup per delivery, so this takes extra work in the model. You collect from the site that’s cheapest today, and sometimes that’s the farther one.
Where pickup and delivery shows up
Medical transport routing optimisation is the clearest case. Specimens and blood are collected from clinics and delivered to a lab, each with a collection window and a deadline after which the sample can no longer be tested. Courier route optimisation always works with pairs. Each parcel is picked up from a sender and delivered to a recipient. One vehicle can often collect several items bound for nearby drops, and a model sees those matches where you can’t.
Furniture delivery with collection is another case. A new sofa goes in, the old one comes out. That means two time windows, one item delivered and one item collected, and a fixed order. Ordinary routing can’t handle that. The same structure appears in spare-part exchanges and in reverse logistics, where goods flow back from customers.
Common questions.
Something else? Ask us directly.
Yes, as long as three things hold. Each item is picked up before it’s delivered, both on the same truck. The load never exceeds capacity at any point along the way. And every stop meets its own time window. A pickup and delivery model checks all three after every stop. So it can mix pickups and deliveries in any order without overloading the truck.
Then the pickup location becomes a decision. The model lists the possible sites and chooses exactly one, together with the route. Picking the nearest site first and routing afterwards is simpler, but it misses the cases where a van drives past a farther site anyway. Standard pickup-and-delivery tools usually can’t do this without custom work.
Keep the regular run as the starting point and let the model add capacity only where it’s needed. The model can prefer the usual vehicle for the usual stops, as long as that van isn’t full. Whatever no longer fits goes to the vehicle that can collect it most cheaply. Drivers keep their routine, and the extra volume still gets picked up.
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”.
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