AI-Powered Design Solutions for MEP

Last year we made significant progress in integrating AI into our design processes and workflows. While there is still much work ahead,  a solid foundation has been established.

Engineering Network Routing Automation

We have focused on developing an advanced tool to partially automate the routing of engineering networks (HVAC, MEP).

• Designers define the start and end points of the route.

• The system automatically calculates the optimal path, accounting for obstacles.

• Provides options for the shortest route or a path that follows walls.

• Supports the inclusion of intermediate points and allows manual adjustments.

• Extensive research into classical routing algorithms.

• Pilot implementation leveraging Reinforcement Learning techniques*.

• 2.5x faster network modeling compared to manual methods.

• Enhanced quality by reducing intersections and conflicts in designs.

*What is Reinforcement Learning Technique?

Reinforcement learning is a method used to train software to make decisions that lead to the most optimal outcomes. It replicates the trial-and-error approach humans use to accomplish their objectives. Actions that align with the desired goal are rewarded and reinforced, while those that move away from the goal are penalized or disregarded.