We turn hard optimisation problems into plans you can run.
NexOR is an optimisation company for industry. We write the mathematics that turns constraints, costs and capacity into a schedule, an allocation or a plan that holds up in production, and we maintain the solvers behind it. Any industry, any problem you can write down.
- Pricing
- Published, per second
- Try it
- In your browser, no account
- Hosted
- In the EU by default
- Built in
- Court-Saint-Étienne, Belgium
See it in 100 seconds
One routing question, the mathematics under it, and the two products we built on it.
The film asks which of the 479,001,600 orders of 12 stops is the cheapest and draws it. It shows the same question in scheduling, rostering, packing, cutting stock and production planning, and every one of them as decisions, constraints and a cost. Then it shows Optimization as a Service, where one model goes to a meta-solver and one answer comes back with its price, and the Transport Management System, from the order to the invoice. It ends on the two ways to start: the Studio and a demo.
- routing
which vehicle goes where, and in what order
- scheduling
what runs when, and on which machine or team
- rostering
who works which shift, within the rules
- packing
what goes in each box, pallet or trailer
- allocation
how to share a limited resource between jobs
- assignment
matching one to one at the lowest cost
- sequencing
the order of jobs that loses the least time
- blending
the cheapest mix that still meets every specification
- cutting stock
how to cut the material with the least waste
- network design
which depots and connections are worth running
- facility location
where to open a site, and how big
- production planning
what to make, when, and on which line
- and yours
the problem that is not on this list
- what you choose
- decision variables
- what must hold
- constraints and bounds
- what it costs
- objective function
Whatever you call it, if you can write those three down, we can solve it.
- linear programming
when amounts vary smoothly and cost is proportional
- mixed-integer programming
when some choices are yes or no, or whole units
- quadratic programming
when doubling the amount more than doubles the cost
- nonlinear programming
when the maths does not follow a straight line
- constraint programming
reasoning from what each choice rules out
- branch and cut
split the problem, cut away what cannot win
- column generation
generate only the options the answer needs
- dynamic programming
solve the stages once, reuse the answers
- metaheuristics
a guided search when proof is out of reach
- local search
improve a plan by changing a little at a time
- shortest paths
the cheapest way from here to there
Two products. Same mathematics underneath.
Optimization as a Service
For engineers and operations researchers. Describe the problem once: routing, scheduling, rostering, packing or allocation. We keep that one model compatible with every engine in the catalogue, so you name one or let a meta-solver pick. Open engines run on our side, so there is no licence to buy first.
- Models
- one model, many solvers
- Billing
- per second of reserved compute
- Start
- free credits, no card
Transport Management System
For dispatchers and fleet owners. A complete transport management system that runs directly in Odoo, on your own database. Give it the day's jobs, vehicles and time windows: orders become a plan, the plan reaches drivers on their phones, and completed work becomes invoices in one batch.
- Route optimisation
- included
- Platform
- your own Odoo
- Pricing
- quoted, no per-driver fees
Milk doesn't wait for a planner. NexOR builds the day's route each morning, pushes it to the drivers, and tells us where every truck is in real time. The history shows us what to fix the next time round. When we ask the team for something, they're on the phone the same day.
Kristof LambrechtsCEO, Transport Lambrechts NV
The name says what we do
Nex from nexus, a binding. OR for operational research. Optimization is what we do with both.
Nex, from nexus
a binding, a connection
Latin, from nectere, to bind. A connection, and the point where several things come together. Here it is literal: decisions, constraints, capacity and costs meet in one model, and the model is what we solve.
OR, for operational research
research, applied to operations
The mathematics of choosing. Out of every plan your constraints allow, it finds the one that costs least to run. We apply it to operations that already exist: shifts, machines, stock and budgets.
Optimization, what we run
a plan you can prove
Not a better plan, a proved one. Cost comes down, the bound that proves it comes up, and the answer carries the gap between them. It runs inside a platform people use every day, and behind an API you call.
Who builds this
Time in the cab, in the lab, and in the codebase.
Between us: years behind a wheel, operational research from the academic side, engineering applied to industry, and a long time on Odoo.
-
Driver Has followed the plan, so knows what it must survive.
-
Mathematician Writes the problem down so a solver can answer it.
-
The academic Finds the case that breaks it, before you do.
-
Researcher Reads what is new, and keeps our methods current.
-
Engineer Turns a model that worked once into software that runs daily.
-
Odoo expert Keeps everything in one database, so nothing needs syncing.
-
Wizard Solves the problem nobody else can even describe.
-
The one who cannot let it go Keeps improving the answer after everyone else stopped.
The people behind them. About NexOR
You write the model. We pick the engine.
The catalogue is large and it keeps growing: engines for linear and mixed-integer programs, constraint models, routing and non-linear problems. Every one is reached through the same interface, so the model you write does not change when the engine does.
A meta-solver is a named set of members, not a black box. It runs your model on a member that can express it, falls through to another when one refuses, and reports which engine actually ran.
Open engines run on our side, so there is nothing to buy first. For a commercial engine you already licence, we use your licence. Every price is published, per second, with no sales call.
The catalogue
No engines are published on this site yet.
Start with the problem you have
You have an optimisation problem
Try it against a real solver, no account and no card. Create an account to keep your work or call the API, and start with free credits.
You run a fleet
See the platform running, and work through how your own operation would fit.
Not sure which
Tell us what you are trying to decide. If it can be written down, we will say which product fits, or which engine. We reply within one business day.