---
title: "Physics-in-the-Loop: A Hybrid Agentic Architecture for Validated CAD Engineering Design"
authors:
  - "Elias Berger"
  - "Muhammad Usama"
  - "Jan Mehlstäubl"
  - "Bernhard Saske"
  - "Kristin Paetzold-Byhain"
doi: "10.24963/ijcai.2026/714"
url: "https://ijcai.org/proceedings/2026/714"
pdf: "https://ijcai.org/proceedings/2026/0714.pdf"
venue: "Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence"
year: 2026
pages: "6417-6425"
track: "AI4Tech: AI Enabling Technologies"
keywords:
  - "Advanced AI4Tech: Generative and LLMs-driven AI4Tech"
  - "Domain-specific AI4Tech: AI4Manufacturing"
  - "AI4T: Domain-specific AI4Tech"
---

# Physics-in-the-Loop: A Hybrid Agentic Architecture for Validated CAD Engineering Design

Elias Berger, Muhammad Usama, Jan Mehlstäubl, Bernhard Saske, Kristin Paetzold-Byhain

*Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence*, pages 6417-6425. DOI: [10.24963/ijcai.2026/714](https://doi.org/10.24963/ijcai.2026/714)

## Abstract

Large Language Models (LLMs) can generate Computer-Aided Design (CAD), yet lack physical comprehension required for reliable engineering design. Instead of attempting to implicitly learn physical laws from data, we propose a Hybrid Agentic-Physical Architecture that embeds validated knowledge-based engineering tools directly into the decision-making loop of autonomous AI agents. In this framework, engineering design is formulated as a closed-loop, sequential decision-making process guided by explicit physical verification. Based on a load case, dedicated agents iteratively plan, generate, evaluate, and revise engineering designs using knowledge-based tools as a feedback signal. We introduce a benchmark dataset and metrics for assessing functional validity in generative CAD. Our system generates more complex and physically verified designs, with a 4.2x increase in structural complexity and improving compile rate by 3.5% compared to similar agentic methods. The codebase, prompts and dataset will be made publicly available to support reproducibility and future research.