Skip to content
Cintrium

Optimisation software for complex operations

We turn operational constraints into software that finds better decisions: routes, rosters, schedules and energy use, planned around the rules your operation actually runs on.

  • Built around your constraints
  • Measured against your current method
  • Full source-code handover

What we optimise

Three kinds of operational decision, each a mathematical problem with real constraints. Forecasting tells you what is likely to happen; optimisation decides what to do about it. We build both, and keep them distinct.

Routes and fleets

Vehicle and crew routes that respect time windows, capacities, skills and shifts, and re-plan when the day changes.

  • Waste and recycling collection
  • Field-service and maintenance crews
  • Delivery fleets
  • Depot and territory allocation
  • Time windows, capacities and shift limits
  • Re-planning for sickness, closures and urgent jobs

Workforce

Rosters and shift plans that cover demand, keep to working-time rules and share the unpopular shifts fairly.

  • Employee rostering
  • Shift allocation to demand
  • Skills and qualifications
  • Leave, absence and disruption
  • Working-time and rest rules
  • Overtime and fairness

Energy and resources

Forecasts of what will happen, and schedules that decide what to do about it, for energy use, storage and assets.

  • Demand and load forecasting
  • Load scheduling and peak reduction
  • Battery and storage dispatch
  • Energy procurement
  • Asset maintenance planning
  • Resource allocation

Standard software handles standard problems

Packaged route planners and scheduling products are built for the common case. When an operation has rules they can't express, the plan they produce is one your staff have to fix by hand. Cintrium builds optimisation systems around your constraints instead.

You get a written model of the operation, a solver that finds the best plan it can prove, scenario tools for what-ifs, and an interface your planners use without writing code. Every result is compared with the strongest simple method we can find and with your current process, and we say how close to the best possible it is proven to be.

The constraints that make a problem yours

  • Several waste streams or job types, each with its own destination
  • Vehicle capacities and weight limits
  • Crew numbers, skills and working patterns
  • Access, timing and one-way restrictions
  • Service times that differ by site
  • Exceptions at individual addresses or assets
  • Manual overrides that must stick
  • Data arriving as CSV, XML or spreadsheet exports

Demonstration systems

Small demonstrations with synthetic data that run live in your browser. They illustrate the kind of systems we design and deliver.

Crew allocation optimisation

Three depots cover six districts. An optimisation model finds the crew allocation with the least total travel, compares it with the rule of thumb it would replace, and re-plans when conditions change.

  1. Inputs: demand per district, crews per depot, travel times.
  2. Model: a minimum-cost flow, solved exactly.
  3. Comparison with the nearest-depot rule.
  4. Scenarios: raise demand or close a depot and re-plan instantly.

What a full project would include

A data model of your assets and constraints, the optimisation engine with tests, a planner interface, an API for other systems, and documentation of every constraint and assumption.

East depot
Travel per crew-hour, optimised
12.6 min
With the nearest-depot rule
13.5 min
Travel time saved
7.3%
Demand left unmet
0crew-hours
Optimised crew-hours sent from each depot to each district, with travel minutes per crew-hour
DepotRiversideOld TownHarbourHillcrestAirportIndustrialCrews used
North
4512 min
020 min
035 min
018 min
040 min
030 min
45 of 120
Central
022 min
708 min
5018 min
4015 min
028 min
020 min
160 of 160
East
038 min
025 min
514 min
032 min
3516 min
6010 min
100 of 100
Demand457055403560

Cells show crew-hours per day and the travel time for each. The nearest-depot rule serves each district in turn from the closest depot with spare crews.

Fuller demonstrations for prospective clients run on real road and asset locations: crew routing across Belgrade's road network, four-week rostering for a 24/7 operation, and a year of maintenance planning for Serbia's high-voltage substations. Ask us for access.

Request access

Typical projects

Each is a defined project with a working system at the end, scoped from a problem statement, a sample of your data and a budget.

Route optimisation system

Plans each day's routes for vehicles and crews under your time windows, capacities, skills and shift rules, with a planner's view and an API.

Workforce scheduling system

Builds rosters that cover every shift, respect working-time rules and requests, and re-solve when someone is off.

Forecasting system

Forecasts of demand, load or volume, tested against simple baselines and fed into the planning models that act on them.

Energy scheduling system

Decides when to run flexible load, charge or discharge storage and buy energy, against forecasts, prices and limits.

Maintenance and asset planning

Chooses which assets to inspect or repair, and when, within budget and crew hours, to reduce expected failure cost.

Scenario-planning tool

Lets planners test a change, such as a closed depot, a budget cut or a new contract, before committing to it.

How we build it

The disciplines behind every system, from the model to the software your planners use.

Mathematical optimisation

  • Mixed-integer programming
  • Vehicle routing
  • Scheduling and rostering
  • Constraint programming
  • Heuristics and local search

Operations research

  • Modelling the operation
  • Baselines and benchmarking
  • Scenario analysis
  • Capacity planning
  • Simulation

Forecasting

  • Demand and load forecasts
  • Backtesting against baselines
  • Uncertainty ranges
  • Weather and calendar effects

Machine learning

  • Prediction models
  • Anomaly detection
  • Feature engineering
  • Model monitoring

Data engineering

  • Data pipelines and validation
  • Integration with existing systems
  • APIs
  • Data migration

Custom software

  • Planner interfaces
  • Web applications
  • Dashboards
  • Deployment and documentation

Build. Deploy. Document. Hand over.

Every project moves through the same five stages, each with a clear output. At the end you own the system and can run it independently.

  1. Define

    Understand the operation, its rules and constraints, the data available and what a better plan is worth.

  2. Prototype

    Build a working model on your real data and compare it with your current method before the full build.

  3. Engineer

    Turn the validated model into production software, with testing, documentation and appropriate security controls.

  4. Deploy

    Deploy into your environment or the infrastructure agreed in the scope.

  5. Handover

    Source code, documentation, deployment instructions and knowledge transfer to your team.

Support and further development can be contracted separately. Neither is a condition of the project.

Built for procurement. Delivered for operators.

Tenders and fixed-scope contracts need a defined scope up front, evidence at each milestone and a clean handover at the end. We run every project that way.

What you receive
DeliverableStage
Defined scope and technical specificationDefine
Written model of the constraints and objectivesDefine
Acceptance criteria, agreed before the buildDefine
Working prototype on your data, compared with your current methodPrototype
Testing evidenceEngineer
Deployment documentationDeploy
Full source codeHandover
System documentation, and training where requiredHandover

Available for direct projects, subcontracting and consortium delivery.

Why Cintrium

A small team of engineers who write the models themselves, direct access to the people building your system, fixed deliverables and nothing you can't take with you.

Technical founders
The people who scope your project are the people who build it. There is no account-management layer in between.
Optimisation at the core
Routing, scheduling and allocation models built by people who write the mathematics, not only configure a product.
Honest benchmarks
Every result is compared with the strongest simple method and with your current process, and how close it is to the best possible is stated, never overclaimed.
Bespoke to your rules
Built for your constraints, data and planners, not a product configured to roughly fit.
Transparent delivery
Scope, milestones and acceptance criteria are written down before the build starts.
No unnecessary lock-in
Open, widely used solvers and technology where they fit, which another competent team could maintain.
Source-code ownership
You receive the full source code and documentation at handover.
Prime supplier or subcontractor
We can deliver a project directly or supply the optimisation component inside a larger programme.

Where these problems come up

Different sectors, similar problems: people, vehicles and assets to plan, rules that can't be broken, and existing tools that don't fit.

  • Local government

    Waste and recycling rounds, highways and grounds crews, care rostering and fleet planning.

  • Energy and utilities

    Demand forecasting, storage and load scheduling, asset maintenance and field crews.

  • Transport and logistics

    Fleet routing, depot allocation, delivery scheduling and capacity planning.

  • Health and care

    Staff rostering, home-visit routing and resource allocation.

  • Field services and facilities

    Crew scheduling, maintenance planning and job allocation.

Security

Security is considered throughout the build, not added at the end. Specific security, regulatory and data-protection requirements are agreed as part of the project scope.

  • Encryption where appropriate
  • Role-based access control
  • Audit logging
  • Least-privilege access
  • Automated testing
  • Dependency management
  • Reproducible deployments
  • Documented infrastructure
  • Data retention controls

A planning problem your software doesn't fit?

Send us the rules, the data you have and what a better plan would be worth. We'll tell you what can be built, how we'd model it and what delivery would involve.

Describe the problem

Discuss a problem

Tell us the operation, the constraints and the budget. If there is a specification or tender pack, link to it in your message.

What happens next

  1. 1We read your enquiry and reply to arrange a short call.
  2. 2We ask the questions needed to understand the operation and scope the work properly.
  3. 3If it's a fit, you receive a written proposal with scope, milestones, acceptance criteria and price.

We use these details only to reply to your enquiry.