Vessel stowage planning
Use ship space while preserving loading and unloading access.
We help the people who run operations make better decisions, with their own data.
Explore five decisions, from source visualizations to recorded simulations. Each case explains the problem and the scope of its evidence.
All applicationsUse ship space while preserving loading and unloading access.
Prioritize waiting retrievals and assign available robot pairs in a simulation.
Compare saved parking layouts, aisles and route-based retrieval estimates.
Review depot locations and service areas through recorded planning iterations.
Coordinate aircraft, gates, crews and ground-service tasks.
Recordings play when fully in view and pause on request. Customer identities are withheld.
Problems we work onLogistics · Manufacturing · Healthcare · Retail · Transport · Defense · Robotics
Ask a question, get an answer: that is where most AI stops. Give the Decision OS a goal and it returns the decision that achieves it. Outcomes and feedback come back as data, so the next decision starts better.
Planners, dispatchers and operators see the recommended plan with its alternatives and reasons. They approve it, override it, or leave a comment, and act on it in the field.
Omelet’s applications are built so that people stay in control of the decision.
Vertical applications for routing, scheduling, loading and dispatch. This is where plans are made, compared and put to work, connected to the systems that execute them.
Every application draws its decisions from the one decision layer below it.
Give it a goal and the rules of the operation. It generates candidate plans, searches for better ones, scores them against real objectives and constraints, and recommends one with its reasons.
This is the layer Omelet builds: the Decision OS. Most AI answers questions; this layer takes a goal and makes the decision. The six capabilities live here.
The meaning of the operation: vehicles, orders, sites, drivers, shifts, and the rules that bind them. One model, shared by every decision.
This is where most AI stops: ask a question, get an accurate answer. Omelet treats it as the input to a decision.
Source data is connected, validated and given context before any decision depends on it, and kept in sync as the operation changes.
Good decisions need trustworthy inputs. This layer earns that trust.
ERP, WMS and TMS records, telematics and sensor events, and what people type in. The loop closes here: outcomes and overrides are captured as data too.
The decision loop closes here: outcomes, overrides and feedback return as data, so the next decision starts better.
Six complementary capabilities, working together.
Connect intent with the right capabilities.
Frame objectives, constraints, and the decision.
Connect people, resources, and changing conditions.
Explore possibilities and evaluate trade-offs.
Reason with context. Explain the result.
Carry reviewed experience into future decisions.
From a question to a reviewed plan, in six steps.
Ask · AI
You ask in your own words. I find the capability that fits.
Frame · Problem-solving intelligence
Pick the objective, then the constraints that must hold.
Ground · Operational intelligence
Bring the operating data. I match the columns and you confirm.
Solve · Optimization
The solver of the OCP Server computes the plan. I do not do the math.
Explain · Inference
I put the result in operational terms, on a map or a table.
Remember · Long-term memory
What you reviewed is kept. The next request starts from it.
What optimization problem would you like to solve?
Optimize one problem end to end, or run and compare multiple scenarios for a more advanced analysis.
Routing
Vehicle routing (TSP, VRP)
Advanced Routing
Routing with several depots and hard or soft constraints
Inventory Control
Replenishment policy per product
Nurse Scheduling
Nurse rostering
Objective
What do you most want to achieve?
Minimize total cost
Balance workload across vehicles
Constraints
Pick anything that must be respected.
Vehicle weight limit
Delivery time windows
Driver working hours
Data
Here is what we have so far.
Upload complete
Uploaded 3 files: depot.csv, vehicles.csv, visits.csv
Column mapping is done - 7 columns, 100% confidence.
Objective function
Minimize total cost100%
Parameters
Computing the plan
Routing · OCP Server · Solver core
1 succeeded / 1 total
Pick an option below to generate a visualization:
1 succeeded / 1 total
Pick an option below to generate a visualization:
Your review
Next time, keep vehicle 2 under four stops.
Record
Saved as a playbook
Carried into the next request
Objective
What do you most want to achieve?
Balance workload
Maximize preference satisfaction
Constraints
Pick anything that must be respected.
Minimum rest between shifts
Guaranteed weekly days off
Maximum weekly hours
Data
Here is what we have so far.
Upload complete
Uploaded 3 files: nurses.csv, shifttypes.csv, forbiddenpatterns.csv
Column mapping is done - 6 columns, 100% confidence.
Objective function
Balance workload100%
Parameters
Computing the plan
Nurse Scheduling · OCP Server · Solver core
1 succeeded / 1 total
Pick an option below to generate a visualization:
1 succeeded / 1 total
Pick an option below to generate a visualization:
Your review
Keep two senior nurses on every night shift.
Record
Saved as a playbook
Carried into the next request
Objective
What do you most want to achieve?
Minimize total inventory cost
Maximize service level
Constraints
Pick anything that must be respected.
Target service level
Minimum order quantity
Shelf life
Data
Here is what we have so far.
Upload complete
Uploaded 2 files: products.csv, demand.csv
Column mapping is done - 5 columns, 100% confidence.
Objective function
Minimize total inventory cost100%
Parameters
Computing the plan
Inventory Control · OCP Server · Solver core
1 succeeded / 1 total
Pick an option below to generate a visualization:
1 succeeded / 1 total
Pick an option below to generate a visualization:
Your review
Review perishables weekly, not monthly.
Record
Saved as a playbook
Carried into the next request
RoutingNurse SchedulingInventory Control
Plan tomorrow’s deliveries. Keep the total cost down.Build next month’s nurse roster. Keep the workload even.Set reorder rules for each product. Keep the total inventory cost down.Describe the problem…
Optimization Results
1 succeeded / 1 total
Routes
3
Stops
10
Unassigned
0
Run interpretation
Three routes cover all ten stops within the weight limit and the delivery windows.
Optimization Results
1 succeeded / 1 total
Nurses
10
Unmet demand
0
Rule violations
0
Run interpretation
Every shift is staffed and no rest rule is broken. Night shifts are spread across the team.
Optimization Results
1 succeeded / 1 total
Products
12
Policies
12
Horizon
30 d
Run interpretation
Each product gets a reorder point and an order quantity that meet the target service level.
Illustrative walkthrough · sample data
Our mission is to solve complex operational problems in industry, turn that problem-solving into reusable assets, and connect more customers’ decisions to their operations. We turn projects into capabilities, and capabilities into products.