Most production control systems are built for static environments. The AI agents act as intelligent process controllers that dynamically fine-tune existing control systems parameters. The agents enable those control systems to adapt to changing manufacturing environment.

Fractal designs the AI agents using manufacturers’ subject matter experts through the Machine Teaching process. The agents are trained using deep reinforcement learning techniques and custom-built AI simulators trained on real-life process data.

The PYO solution combines accelerators, best practices, and custom data science engagements to build, train, and deploy effective AI agents.

The solution includes:

• Initial project assessment

• Data preparation and integration as required to enable AI at the edge.

• Accelerators for end-to-end design, simulation development, agent training, agent testing, and deployment at the edge.

• Development of a custom simulation leveraging the most appropriate approach for the customer process as part of DRL deployment: third-party simulation tools (e.g., AnyLogic, Simulink, etc.), custom physics-based models, custom digital twin-based simulations, or trained AI simulations for DRL purchases.

• Field deployment using Azure IoT Edge Runtime, AKS, Kubernetes, or other edge container services depending on the skill.

• Field testing on real-life processes & rollout deployment assistance.

Note: This solution does not include wiring, custom electronics work, or device deployment services.

Depending on its complexity and stage, different experts will be brought in throughout the initial 3-6 month deployment. Autonomous AI projects using Deep Reinforcement Learning DRL) often take longer, typically 6-12 months, before being fully deployed and running. The minimum commitment to purchase is 1 year, but it is expected our management services will continue ongoing for all deployments, with constant improvements through our managed service upgrades leveraging the latest Azure technologies.

This solution has been validated for Azure private MEC for edge connectivity and compute capability. It also has been validated on Azure Stack Edge and Azure HCI.