EDA design intelligent system and method based on multi-agent cooperation

By using a multi-agent collaborative system and a directed graph feedback mechanism, the problems of long design cycles and poor accuracy in EDA design are solved. It achieves efficient and accurate conversion from natural language requirements to hardware implementation, thereby improving the robustness and applicability of the design process.

CN122113769APending Publication Date: 2026-05-29HANGZHOU INNOVATION RES INST OF BEIJING UNIV OF AERONAUTICS & ASTRONAUTICS +1
0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
HANGZHOU INNOVATION RES INST OF BEIJING UNIV OF AERONAUTICS & ASTRONAUTICS
Filing Date
2026-02-04
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

Existing EDA design systems suffer from long design cycles, high costs, heavy reliance on engineers' experience, and a lack of closed-loop feedback and multi-dimensional task collaboration capabilities, making it difficult to achieve efficient and accurate conversion from natural language requirements to hardware implementation.

Method used

A multi-agent collaborative system is adopted, including an EDA design knowledge agent, a task planning agent, an HDL code generation agent, and multiple EDA tool agents. A directed graph feedback mechanism is used to achieve closed-loop iterative optimization of the design process. Combined with large model fine-tuning data, manual intervention is supported to deal with design anomalies.

Benefits of technology

Significantly shortens the design cycle, improves design accuracy and efficiency, reduces the frequency of manual intervention, achieves a PPA (Progressive Product Achievement) compliance rate of over 95%, and enhances the robustness and applicability of the design process.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122113769A_ABST
    Figure CN122113769A_ABST
Patent Text Reader

Abstract

The present application relates to a kind of EDA design intelligent system and method based on multi-agent cooperation.The collaborative system including EDA design knowledge agent, task planning agent, HDL code generation agent and multiple EDA tool agent is constructed.Each agent realizes the closed loop process of demand analysis, HDL generation, tool execution, result feedback and iterative optimization through directed graph feedback mechanism, real-time sensing EDA tool execution result and dynamically generating optimization strategy.The present application does not need to rely on external case database, can automatically analyze PPA deviation root cause and trigger targeted optimization, while realizing man-machine cooperation by artificial intervention mechanism.The system significantly shortens design cycle, improves PPA compliance rate, reduces artificial intervention frequency, realizes from natural language demand to high PPA hardware implementation End-to-end automation, provides efficient, accurate, low-reliance solution for chip design intelligence.
Need to check novelty before this filing date? Find Prior Art