Home // DATA ANALYTICS 2025, The Fourteenth International Conference on Data Analytics // View article

Parametric Optimization and Intelligent CAD Automation for Bicycle Frame Design via Multi Agent

Authors:
Chon Kit Chan
Wen-Yi Yang
Jin H. Huang

Keywords: Bicycle rim design; large language model; retrieval-augmented generation; multi-agent system; intelligent CAD automation.

Abstract:
This research presents an intelligent design platform integrating Large Language Models (LLMs), Retrieval Augmented Generation (RAG), a multi-agent system, and the Model Context Protocol (MCP). It addresses the inefficiencies, high costs, and heavy reliance on manual expertise in traditional bicycle design workflows, which constrain both innovation and responsiveness to market demands. This challenge is especially pressing as the bicycle industry faces increasing demands for customization, small-batch production, and rapid product development. While prior approaches such as Machine learning(ML) and data-driven optimization have shown promise, they remain confined to isolated tasks and lack a unified, end-to-end framework. The proposed system leverages these technologies to generate context-aware design recommendations tailored to user intent and automates Computer Aided Design (CAD) model generation, thereby substantially reducing development time and manual workload. As a proof of concept, the platform is first applied to rim design the component with the richest dataset before scaling to other components and full frame development. Experimental results demonstrate that the system can reduce design cycles from weeks to hours, providing higher efficiency, lower costs, and improved accuracy of design recommendations. The goal is to accelerate intelligent, flexible, and automated design workflows for the next generation of bicycle products.

Pages: 7 to 14

Copyright: Copyright (c) IARIA, 2025

Publication date: September 28, 2025

Published in: conference

ISSN: 2308-4464

ISBN: 978-1-68558-293-7

Location: Lisbon, Portugal

Dates: from September 28, 2025 to October 2, 2025