Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Multi-Agent Collaboration for Automated Design Exploration on High Performance Computing Systems

Harshitha Menon, Charles F. Jekel, Kevin Korner, Brian Gunnarson, Nathan K. Brown, Michael Stees, M. Giselle Fernandez-Godino, Walter Nissen, Meir H. Shachar, Dane M. Sterbentz, William J. Schill, Yue Hao, Robert Rieben, William Quadros, Steve Owen, Scott Mitchell, Ismael D. Boureima, Jonathan L. Belof

TL;DR

This work presents MADA (Multi-Agent Design Assistant), a Large Language Model (LLM) powered multi-agent framework that coordinates specialized agents for complex design workflows that reduces cumbersome manual workflow setup, and enables automated design exploration at scale.

Abstract

Today's scientific challenges, from climate modeling to Inertial Confinement Fusion design to novel material design, require exploring huge design spaces. In order to enable high-impact scientific discovery, we need to scale up our ability to test hypotheses, generate results, and learn from them rapidly. We present MADA (Multi-Agent Design Assistant), a Large Language Model (LLM) powered multi-agent framework that coordinates specialized agents for complex design workflows. A Job Management Agent (JMA) launches and manages ensemble simulations on HPC systems, a Geometry Agent (GA) generates meshes, and an Inverse Design Agent (IDA) proposes new designs informed by simulation outcomes. While general purpose, we focus development and validation on Richtmyer--Meshkov Instability (RMI) suppression, a critical challenge in Inertial Confinement Fusion. We evaluate on two complementary settings: running a hydrodynamics simulations on HPC systems, and using a pre-trained machine learning surrogate for rapid design exploration. Our results demonstrate that the MADA system successfully executes iterative design refinement, automatically improving designs toward optimal RMI suppression with minimal manual intervention. Our framework reduces cumbersome manual workflow setup, and enables automated design exploration at scale. More broadly, it demonstrates a reusable pattern for coupling reasoning, simulation, specialized tools, and coordinated workflows to accelerate scientific discovery.

Multi-Agent Collaboration for Automated Design Exploration on High Performance Computing Systems

TL;DR

This work presents MADA (Multi-Agent Design Assistant), a Large Language Model (LLM) powered multi-agent framework that coordinates specialized agents for complex design workflows that reduces cumbersome manual workflow setup, and enables automated design exploration at scale.

Abstract

Today's scientific challenges, from climate modeling to Inertial Confinement Fusion design to novel material design, require exploring huge design spaces. In order to enable high-impact scientific discovery, we need to scale up our ability to test hypotheses, generate results, and learn from them rapidly. We present MADA (Multi-Agent Design Assistant), a Large Language Model (LLM) powered multi-agent framework that coordinates specialized agents for complex design workflows. A Job Management Agent (JMA) launches and manages ensemble simulations on HPC systems, a Geometry Agent (GA) generates meshes, and an Inverse Design Agent (IDA) proposes new designs informed by simulation outcomes. While general purpose, we focus development and validation on Richtmyer--Meshkov Instability (RMI) suppression, a critical challenge in Inertial Confinement Fusion. We evaluate on two complementary settings: running a hydrodynamics simulations on HPC systems, and using a pre-trained machine learning surrogate for rapid design exploration. Our results demonstrate that the MADA system successfully executes iterative design refinement, automatically improving designs toward optimal RMI suppression with minimal manual intervention. Our framework reduces cumbersome manual workflow setup, and enables automated design exploration at scale. More broadly, it demonstrates a reusable pattern for coupling reasoning, simulation, specialized tools, and coordinated workflows to accelerate scientific discovery.
Paper Structure (39 sections, 8 figures)

This paper contains 39 sections, 8 figures.

Table of Contents

  1. Introduction
  2. Background
  3. Agents and Tool Use
  4. Multi-Agent Systems
  5. Model Context Protocol (MCP)
  6. MADA System Overview
  7. Specialized Agents
  8. MADA Orchestration and Agent Coordination
  9. Tool Integration via MCP
  10. Job Management Agent (JMA)
  11. Simulation Code Interface
  12. Laghos
  13. Scheduler Integration
  14. Geometry Agent (GA)
  15. Interacting with Cubit
  16. ...and 24 more sections

Figures (8)

  • Figure 1: MADA system architecture. The framework orchestrates three specialized agents: the Job Management Agent (JMA) manages simulations on HPC via Flux, the Geometry Agent (GA) generates meshes through Cubit, and the Inverse Design Agent (IDA) explores the design space. Agents communicate with tools via the Model Context Protocol (MCP).
  • Figure 2: Coordination process. The context analyzer processes conversation history and agent roles. The selector chooses the next speaker. The selected agent responds, and the system broadcasts the message to maintain shared state. This cycle repeats until the system reaches termination.
  • Figure 3: Simulation setup of inertial confinement fusion in 2D with sinusoidal energy initialization in the x-axis and a perturbed interface to trigger RMI korner2025differentiable.
  • Figure 4: Design exploration results for RMI suppression. Top row: initial energy field. Bottom row: final density profile. Left: configuration at the start of exploration (QoI = 4.1). Right: best design found by MADA (QoI = 3.7), showing significantly reduced jetting at the interface.
  • Figure 5: Visualization of the high-velocity impact problem that results in the formation of RMI from 10.1063/5.0100100.
  • ...and 3 more figures