Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study introduces a lightweight multi-agent framework for the automated design of reinforced concrete highway barriers, addressing the limitations of current manual and heuristic methods. By employing a "generation-evaluation-optimization" closed-loop approach, the framework achieves over 98% design accuracy, outperforming larger LLMs while demonstrating that model performance does not scale with size. The findings suggest significant reductions in computational costs and improved accessibility for AI-assisted engineering tools in industry applications.
An 8B-parameter model outperforms a 631B-parameter flagship model in concrete barrier design, challenging assumptions about model size and performance.
The design of reinforced concrete highway barriers is a safety-critical process that requires strict compliance with regulatory provisions such as the AASHTO-LRFD bridge design guidelines. Current engineering practice relies heavily on manual, iterative, and heuristic calculations to satisfy complex nonlinear material and mechanics constraints. Although Large Language Models (LLMs) demonstrate strong generative capabilities, their direct application to structural engineering remains limited by hallucination risks and insufficient physical grounding. To address these challenges, this study proposes a novel "generation-evaluation-optimization" closed-loop framework for automated concrete barrier design using the multi-agent orchestration capabilities of AutoGen. Experimental results demonstrate that the proposed agentic framework achieves over 98% design accuracy, significantly outperforming standalone general-purpose LLMs. More importantly, the study reveals that design performance is not necessarily correlated with model scale, where an 8B-parameter lightweight model could outperform unconstrained 631B-parameter flagship models. This finding highlights the potential to substantially reduce computational costs while improving the accessibility of AI-assisted engineering tools for industry applications. The source code for the proposed multi-agent design framework is available at the project GitHub repository: https://github.com/MXY820/barrier-design. Keywords: Structural Engineering; Multi-Agent Systems; Large Language Models; Concrete Barrier Design; AutoGen; Design Automation.