ODYSSEE
Real-time engineering problem solving
ODYSSEE is a powerful innovation platform that allows engineers to apply machine learning, Artificial Intelligence (AI), Reduced Order Modelling (ROM) and design optimisation to workflows.
Benefit from modern data science technology by creating cost efficient digital twins through real-time modeling and optimization for both CAE simulation data and physical test data.
ODYSSEE includes:
- Machine Learning & AI
- Statistics, Data Mining, Data Fusion
- Optimization and Robustness
- Process Discovery
- Image Recognition and Compression
ODYSSEE is available in two versions. ODYSSEE A-Eye and ODYSSEE CAE.
ODYSSEE CAE
ODYSSEE CAE enhances your knowledge by answering complex engineering questions in real-time that would otherwise take hundreds of hours to simulate and analyze. With only a few previous CAE simulations, ODYSSEE predicts, optimizes, and robustly generates accurate results in real-time. ODYSSEE CAE produces full-time history outputs, including complete CAE analysis with detailed post-processing of results.
ODYSSEE CAE is applicable to all design and engineering applications and is both software and physics independent. It employs algebraic or machine learning solutions for reducing the volume of data while preserving the most important parts of the information contained within the data. This is commonly done via decomposition or machine learning or other efficient data fusion techniques. Such techniques allow for creating on-board and real-time applications based on existing experimental or simulation results. Typical applications are optimization, parametric sensitivity analysis and robustness.
Rethinking Design COST (Computing, Optimization, Simulation and Time)
Computing
ODYSSEE CAE reduces the effective number of CPU intensive computations, replacing them with real-time equivalents capable of being run on small size laptops.
Optimization
Combining optimization and learning to provide very precise surrogate models that can
be employed at very low computing cost allows ODYSSEE CAE to perform very robust optimization by allowing for thousands of runs in seconds or minutes.
Simulation
ODYSSEE CAE establishes the missing link, combining learning from experiments and results from simulations. This improves the performance, the precision, and the feasibility of simulation-based engineering.
Time
ODYSSEE CAE is based on real-time computing; therefore, you need little-computing effort for parametric studies and optimization, exploring new horizons in data science to capture the most important part of the solution, leaving the details to more elaborate and CPU intensive computing, optimization, and simulation technology.
ODYSSEE A-Eye
ODYSSEE A-Eye allows you to create your own AI application based on image data, sensor data, scalars, labels, curves, and CAD data as inputs and then predicts responses using the known data of your system. This insight enables designers and production technicians to explore the design space more extensively and interactively and improve next-generation products without prohibitive computing cost or time.
You can customize your AI applications using the ODYSSEE A-Eye engine by defining the descriptive variables and results. Once the A-Eye engine has learned the problem, it can then predict the result in real-time for new descriptive inputs.
ODYSSEE A-Eye can also perform intelligent data compression on images (best possible compression rate), while keeping as much as possible the useful information.
Contact us
Please contact us for more information about ODYSSEE.
Real-time engineering intelligence — powered by AI, built for engineers.
ODYSSEE transforms your existing simulation data into live, interactive design tools. Machine Learning, Reduced Order Modelling and AI converge in a platform that answers complex engineering questions in seconds — not hours — on a standard laptop.
Watch ODYSSEE Think
A Reduced Order Model predicting stress distribution in real-time as design parameters change — on a standard laptop, no HPC required.
ODYSSEE is Cadence's AI and machine learning platform built specifically for engineers. Unlike generic ML frameworks, ODYSSEE is engineered from the ground up to work with the small, high-quality datasets that characterise CAE simulation environments — where generating data is expensive and accuracy is non-negotiable.
At its core, ODYSSEE uses Reduced Order Modelling (ROM) — a technique that compresses high-fidelity simulation knowledge into a compact, instantly-running model. Once trained on your existing simulation data, the ROM delivers results of comparable accuracy in real-time, on standard hardware, accessible to anyone on your team.
The result: your best engineers' simulation knowledge, encoded as an accessible, interactive tool that the whole team can use — without waiting hours for solver runs or requiring deep AI expertise.
- Works with as little as 10% of data required by traditional ML
- Real-time full-field predictions — stress, deformation, temperature, flow
- No deep ML expertise required — designed for simulation engineers
- Runs on standard laptop hardware after ROM training
- Integrates with Nastran, Adams, Marc, Cradle, Simufact and more
- Multiphysics ROM fusion — combine FEA, CFD and MBD in one model
Key Facts
| Developer | Cadence (MSC Software) |
| Technology | AI / ML / Reduced Order Modelling |
| Training input | CAE results, test data, images, sensor data, CAD |
| Run-time output | Real-time field results, predictions, optimisation |
| Hardware | Standard laptop — no HPC required at run-time |
| AI engine | Quasar — powers all ODYSSEE products |
| Data requirement | ∼10% of traditional ML data volume |
See ODYSSEE Live
We can demonstrate ODYSSEE running on engineering challenges similar to yours — including your own simulation data types.
Proven Results. Real Engineering. Real Numbers.
From global Tier 1 suppliers and OEMs using ODYSSEE in production.
The flagship product. ODYSSEE CAE trains a Reduced Order Model on your existing FEA, CFD, MBD or crash simulation data. Once trained, the ROM runs in real-time on a standard laptop — delivering full-field results across thousands of design variants in the time a single traditional solver run would take.
Design space exploration, robustness studies and multi-objective optimisation become genuinely interactive. Engineers can vary dozens of parameters simultaneously and see deformation maps, stress contours and temperature fields update instantly.
- ROM creation from any solver output (FEA, CFD, MBD, crash)
- Real-time full-field predictions: stress, deformation, temperature
- Multi-objective optimisation in seconds
- Robustness and sensitivity studies
- Multiphysics ROM fusion (FEA + CFD + MBD combined)
- Uses existing simulation data — no new solver runs required
- Runs on laptop — frees up your HPC cluster
ODYSSEE A-Eye brings image-based machine learning to engineering. It learns to predict engineering responses directly from geometry — no meshing, no model setup, no solver runs. Feed it images, CAD geometry, sensor data or scalars, and it learns the relationship between shape and performance.
Ideal for early-stage design screening where geometry is changing rapidly and setting up full FEA models for every variant is impractical. A-Eye democratises simulation access to designers who have never opened a solver.
- AI predictions directly from CAD images and geometry
- No meshing or model preparation required
- Accepts images, CAD, sensor data, scalars, curves as inputs
- Predicts responses: deformation, stress, temperature, pressure
- Intelligent image compression preserving engineering content
- Continuous learning — model refines as new data arrives
- Structural, thermal, fluid, acoustic and crash applications
ODYSSEE Explore
ODYSSEE Explore is the data analytics and visualisation layer of the platform. It lets engineers investigate large simulation databases, identify trends, cluster results, and understand the relationships between design parameters and performance metrics — without writing a single line of code.
Turn years of accumulated simulation runs into actionable engineering knowledge. Find the patterns in your data that no individual engineer has ever had time to look for.
- Visual exploration of large CAE simulation databases
- Statistical analysis, data mining and data fusion
- Parameter correlation and sensitivity mapping
- Outlier detection and quality screening
- Interactive dashboards — no coding required
- Combine simulation results with physical test data
- Cluster analysis to identify design families
ODYSSEE Solver — Quasar
Quasar is the AI engine that powers all ODYSSEE products. It is also available as a standalone development environment, enabling teams to build custom AI applications tailored to specific simulation types or proprietary internal workflows.
Unlike TensorFlow or PyTorch, Quasar is designed from the ground up for engineering datasets — making AI practical for teams who cannot afford to generate millions of training samples.
- Object-oriented AI programming designed for CAE engineers
- Python support for scripting and workflow automation
- Natively embedded in MSC Nastran, Adams, Marc and Cradle
- Build and deploy custom AI models for any engineering domain
- FMU export for real-time co-simulation with Adams and Simulink
- Enterprise API for integration with existing CAE workflows
- Scales from single engineer to enterprise deployment
From your existing simulation data to real-time engineering intelligence — in four steps.
Upload your existing simulation results, physical test data, or CAD images. No new solver runs needed — ODYSSEE learns from what you already have.
ODYSSEE builds a Reduced Order Model using the Quasar AI engine. The ROM captures the engineering physics in a compact, fast-running representation.
Run parametric studies, optimisation loops and robustness analyses in real-time. Thousands of design variants evaluated in minutes — on a laptop.
Share the ROM with your team. Any engineer can query it and receive accurate full-field results instantly. The model improves automatically as new data arrives.
ODYSSEE accelerates simulation across every major engineering physics domain.
Impact prediction, seat crash simulation, interior design, payload analysis. Faurecia achieved 97% time reduction on crash seating simulations for the Cockpit of the Future.
Real-time thermal field prediction for electronics cooling, powertrain and industrial processes. Explore hundreds of thermal variants per hour on a laptop.
Flow analysis, transient modelling, pump and turbine optimisation. Operational CFD Digital Twins updated with real-time boundary condition data from sensors.
Chassis optimisation, racing performance prediction, ROM from Adams models. Export as FMU for real-time embedding in control system co-simulation.
Structural life prediction and buckling analysis in real-time. Explore load combinations interactively without triggering full FEA runs for every variant.
AM process performance, material characterisation, welding distortion prediction. Accelerate Simufact workflows with AI-driven parameter space exploration.
Flight simulation, structural certification studies, aeroelastic screening. Run compliance design-space sweeps at design speed — not certification speed.
Satellite and MRI image compression, sensor classification, CAD-based geometry-to-performance prediction. ODYSSEE A-Eye applies ROM principles to visual and signal data.
The same engineering question. A fundamentally different answer time.
| Capability | Traditional CAE | ODYSSEE + ROM |
| Single analysis run time | Hours to days | Seconds |
| Design space exploration (1,000 variants) | Weeks of HPC time | Minutes on a laptop |
| Hardware requirement | HPC cluster or cloud | Standard laptop |
| Who can run the analysis | Senior analyst only | Any engineer on the team |
| Multiphysics coupling | Complex, expensive, slow | Single ROM fuses FEA + CFD + MBD |
| Sensitivity and robustness studies | Impractical at scale | Interactive, real-time |
| Geometry-to-prediction (no meshing) | Not possible | ODYSSEE A-Eye |
Global manufacturers and Tier 1 suppliers using ODYSSEE in production workflows.
Every ODYSSEE product is built on Quasar — Cadence's proprietary AI engine developed specifically for engineering applications. Quasar is not a generic ML framework: it is designed to work with the small, structured, physically-meaningful datasets that characterise engineering simulation.
Quasar is embedded across the Cadence portfolio, meaning ODYSSEE-trained models operate within the tools your team already uses — not as an external AI add-on, but as native intelligence inside your simulation environment.
- Native AI in MSC Nastran (smart sub-modelling acceleration)
- Embedded in MSC Adams via FMU co-simulation export
- Real-time CFD digital twins via Cradle integration
- Marc nonlinear analysis AI-driven acceleration
- Python API for scripting and workflow automation
- Enterprise REST API for system integration
⚡ The Quasar AI Engine
An object-oriented AI programming language and runtime built for CAE. Quasar manages the mathematics of ROM creation, optimises model accuracy, and exposes a clean programming interface for custom engineering AI applications.
Unlike TensorFlow or PyTorch, Quasar works efficiently with small engineering datasets — making AI practical for simulation teams who cannot afford to generate millions of training samples. As little as a few dozen high-quality simulation runs can produce a useful ROM.
ODYSSEE trains on data from any solver and delivers ROMs that slot into your existing workflow.
ODYSSEE can ingest results from any solver that produces structured output data. The AI engine is solver-agnostic: if you can export results, ODYSSEE can learn from them.
Ready to make your simulations real-time?
Book a live demonstration — we'll show ODYSSEE running on engineering challenges similar to yours.
Frequently Asked Questions
ODYSSEE is Cadence's AI and machine learning platform for engineers, built around Reduced Order Modelling (ROM) — a technique that compresses high-fidelity simulation knowledge into a compact model that runs in real time.
ODYSSEE is designed to work with the small, high-quality datasets typical of CAE environments, and can work with as little as around 10% of the data volume that traditional machine learning approaches require.
No — ODYSSEE is engineered specifically for simulation engineers, not data scientists, and does not require deep ML expertise to train or use a model.
Once a Reduced Order Model is trained, it runs in real time on standard laptop hardware — no HPC cluster is required at run-time.
ODYSSEE can build reduced order models from data produced by MSC Nastran, Adams, Marc, Cradle, Simufact and other CAE tools, and can fuse multiple physics (FEA, CFD, MBD) into a single model.