Robot fleets and AGVs
Fleet size, lane layout, station count and live traffic control for warehouse and factory robots.
Will 200 robots meet our peak, or do we need 300?
Quantum-inspired optimisation
Better decisions for industry, found by quantum-inspired solvers.
We turn floor plans, production schedules and logistics networks into optimisation problems, and solve them with the best available solvers on GPUs and FPGAs. Every plan is proven in simulation before it reaches the floor.
LIVE · annealing on a 36-stop route
–
Quantum-inspired
Quantum-inspired algorithms that run on today’s hardware.
GPU and FPGA
Large design searches, and millisecond decisions on site.
Proven in simulation
Every recommended plan is shown running before you invest.
Open and secure
Open standards, on-premise or Indian cloud, your data stays yours.
Solutions
Problems with more options than anyone can test
A warehouse with 300 robots or a plant with 40 jobs on 10 machines has more possible plans than any computer could ever list one by one. Rules of thumb find a plan that works; we find the best one and show what it is worth.
Production scheduling
Job sequences, changeovers and shift patterns that meet promise dates with the machines you have.
Can we take this order and still ship on time?
Layout and line balancing
Station counts, buffer sizes and work content, and where the bottleneck moves as demand grows.
Which line design reaches the target cycle time?
Logistics and routing
Pick routes, milk runs, tugger loops, dock slots and delivery routes across sites.
How many tuggers, on which loops?
Aerospace manufacturing
Assembly-line balancing, autoclave batch loading for composites, hangar and maintenance slots.
Which parts cure together in this autoclave run?
Energy-aware scheduling
Furnace, melt and heat-treatment sequences under a power limit or a time-of-day tariff.
Which sequence keeps us under contract demand?
Technology
Quantum-inspired solvers on today’s hardware
Quantum-inspired algorithms borrow the mathematics of quantum devices, such as oscillators that settle into the best answer, and run it on CPUs, GPUs and FPGAs. No quantum computer needed.
- Exact and constraint solvers prove plans optimal on small and medium problems.
- Quantum-inspired algorithms find near-optimal plans fast on large problems, in parallel by design.
- FPGA deployment for decisions needed within milliseconds on site, such as live robot traffic (in development).
We benchmark every solver on your problem and use the one that wins. Sometimes that is a classical solver, and we will say so.
Quantum-inspired solver · 20-node max-cut cut – · optimum –
Results
What optimisation changes
From our own prototypes, simulated on public benchmark data. Customer results will replace these as pilots complete.
Warehouse robot fleet · 5,000 deliveries/h target
300robots meet the target with optimised traffic
255/hbest with two-way aisles: gridlock
216designs simulated in one search
Best throughput reached, any fleet up to 400 robots (deliveries/h)
| Design | Deliveries per hour |
|---|---|
| Two-way aisles, any controller | 255 |
| One-way aisles, optimised traffic | about 2,890 |
| Street grid, rule-based traffic | 3,030 |
| Street grid, optimised traffic | 6,582 |
Production scheduling · 12 jobs, 5 machines finish – → –
The same search reorders jobs to cut idle machine time and finish sooner. Real plants add changeovers, due dates and shifts as rules in the same model. Hover a bar to follow one job.
Approach
From your data to a plan you can see running
- 01
Share your data
Floor plan (DXF or PDF), routings, cycle times, order history and the target you need to hit.
- 02
Model the problem
Site or plant model, constraints and goals, checked with your team.
- 03
Search the options
Solvers score thousands of designs: layouts, fleet sizes, sequences, station counts.
- 04
Prove in simulation
The best plans run in a plant simulation: throughput, lead time and bottlenecks, before you spend.
- 05
Decide and deploy
Plans are recommendations. We integrate through open standards such as VDA 5050.
Numbers you can check
Every figure traces to your data or a simulation run you can repeat.
People decide
We recommend plans; we do not write to your machines.
No lock-in
Open-source solvers, DXF, VDA 5050. Your models stay usable without us.
Your data stays yours
On-premise or Indian cloud region, under NDA, never reused.
Quantum-ready
Value today. Ready for quantum hardware tomorrow.
Benefits today
Better plans on current hardware
Quantum-inspired solvers on GPUs and FPGAs deliver results now, with no quantum computer in the loop.
Hardware-ready models
Formulated as QUBO and Ising problems
The same problem formulations that quantum annealers and gate-based machines run, so nothing is rebuilt later.
Quantum advantage tomorrow
Switch when it pays
When quantum hardware beats classical on a problem class, your models move to it. Until then, we use whatever wins.
Partners
Built with partners, not in isolation
Optimisation is only useful when it is proven on the real plant. We work with specialist partners to model, simulate and deliver.
Simulation partner
Apexiz
An industrial simulation team in Hyderabad: plant simulation, layout validation and throughput analysis, robot offline programming and virtual commissioning.
apexiz.comFAQ
Common questions
Do I need a quantum computer?
No. Quantum-inspired algorithms run on ordinary servers, GPUs and FPGAs, on-premise or in the cloud.
Are quantum-inspired solvers always better?
No. On many problems a classical solver is as good or better. We benchmark the options on your problem and use the one that gives the best plan in the time available.
What do you need from us to start?
A floor plan or line layout, the process data you already have and the target you need to meet. We agree scope and data handling before any data moves.
Do you replace our fleet manager, MES or ERP?
No. We plan and simulate; results go to your existing systems through open interfaces.
Are the animations on this page real?
Yes. Each runs the algorithm it describes in your browser on a randomly generated problem. Warehouse figures come from our simulated prototype.
Have a decision with too many options?
Tell us what you need to decide and what data you have. We will reply with how we would approach it.
sitaram@kernallogicautomation.com