Unique AI-solution

Calejo has developed a patent-pending next generation simulation and optimization engine specially designed for artificial intelligence and used in areas such as vehicles, smart cities and industrial processes. We have named it Calejo Optimize. It delivers high performance scalable simulations of large non-linear dynamical systems with millions of equations and variables. Unlike previous generation of simulation engines, it allows for the inclusion of effective automatic model adaptation from sensor data using supervised approaches, which sets all parameters to values that correspond to their real physical properties without need for human adjustments. It allows interpretable and reusable black box/grey box modelling of unknown dynamics and deep learning/neural networks, including compatibility with recent industry standards such as ONNX. It is also compatible with Simulink or Modelica-like graphical user interfaces.


Calejo Optimize also allows control optimization using an entire new generation of reinforcement learning algorithms trained in such environments to achieve desired objective functions, such as balances between energy efficiency and product/service quality. Calejos optimization methods have entirely different and improved algorithmic scaling properties than conventional policy gradient or temporal difference-based methods, which allows automatically created non-linear control of any complexity. The optimized control policies are easily and flexibly converted into efficient code for implementation in real-time control systems and edge computing.

On a PC or on the Cloud

Calejos hierarchical knowledge representation also allows efficient data reuse and knowledge extraction for use in the planning of new blueprints and construction plans, with optimal control for any hypothetical future system developed on the fly from the schematics.  The engine is lightweight and currently runs on any laptop (Win, Linux, Mac, edge etc), but is also already prepared for GPU computing and distributed deployment on the cloud when the need arises.