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

Harshitha Menon; Charles F. Jekel; Kevin Korner; M. Giselle Fernandez-Godino; Brian Gunnarson; Nathan K. Brown; Michael Stees; Walter Nissen

cs.AIMar 12, 2026

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

Authors: Harshitha Menon, Charles F. Jekel, Kevin Korner, M. Giselle Fernandez-Godino, Brian Gunnarson, Nathan K. Brown, Michael Stees, Walter Nissen, +10 more

Organizations: Lawrence Livermore National Laboratory, Livermore, CA, USA · Sandia National Laboratories, Albuquerque, NM, USA · Los Alamos National Laboratory, Los Alamos, NM, USA · AMD

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.

Figures & tables

Fig. 1: MADA system architecture. The framework orchestrates three specialized agents: the JMA manages simulations on HPC via Flux, the GA generates meshes through Cubit or PMesh, and the IDA explores the design space. Agents communicate with tools via MCP.

Server	Function	Description
Flux	submit_job	Single job submission
Flux	submit_jobs_async	Async ensemble submission
Flux	check_job_status	Query job state
Laghos	generate_runs	Stage simulation inputs
Laghos	get_qoi	Extract jet length
MARBL	prepare_parameter_sweep	Generate sweep dirs & inputs

TABLE I: MCP servers and the functions they expose to agents.

Fig. 2: Simulation setup for 2D RMI [ 41 ] . Top: mesh with perturbed interface generated by the GA. Bottom: energy initialization defined by the sampled design parameters.

Fig. 3: 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.

Study	Runs	Best $Q_{\mathrm{mix}}$	Purpose
Initial	200	0.0244	Global 10-variable sweep
0170 main	1000	0.0202	Refine around 0170 with modes $(4,3)$
0170 crosscheck	120	0.0329	0170’s neighborhood with 0166’s modes $(6,2)$
0166 main	420	0.0204	Refine around 0166 with modes $(6,2)$
0166 crosscheck	60	0.0179	0166’s neighborhood with 0170’s modes $(4,3)$

TABLE II: RTI campaign structure and best observed results per study.

Explore similar work

MADDIE: Multi-Agent Drug Discovery Engine
Oct 12, 2026Ye Liu, Reza Averly, Avery Meyer +4Drug Discovery Multi-Agent LLM Systems
Biomedical Informatics, The Ohio State University. · Computer Science and Engineering, The Ohio State University. · Medicinal Chemistry and Pharmacognosy, The Ohio State University. +1
Multi-Agent Orchestration for High-Throughput Materials Screening on a Leadership-Class System
Apr 9, 2026Thang Duc Pham, Harikrishna Tummalapalli, Fakhrul Hasan Bhuiyan +5Multi-Agent Orchestration Scientific Workflow Automation
Computational Science (CPS) Division, Argonne National Laboratory, Lemont IL 60439, USA · Argonne Leadership Computing Facility (ALCF), Argonne National Laboratory, Lemont IL 60439, USA
MDForge: Agentic Molecular Dynamics Pipeline Design under Sparse Simulator Feedback
Jun 11, 2026Zehong Wang, Yijun Ma, Connor R. Schmidt +7AI Agents for Scientific Discovery Molecular Dynamics Simulation
University of Notre Dame · University of Connecticut

Oct 12, 2026J/K move · Enter open · S save

MADDIE: Multi-Agent Drug Discovery Engine

Ye Liu, Reza Averly, Avery Meyer, Frazier N. Baker+3

Biomedical Informatics, The Ohio State University. · Computer Science and Engineering, The Ohio State University. · Medicinal Chemistry and Pharmacognosy, The Ohio State University.+1

Drug Discovery Multi-Agent LLM Systems LLM Agent Orchestration

Open paper AskPDF arXiv

Drug discovery remains a resource-intensive and time-consuming process, involving billions of dollars and a decade of development to advance a drug to market. While many in silico tools have accelerated this process, operating them still demands substantial time, effort, and expertise from medicinal chemists. Autonomous systems built on Large Language Models (LLMs) have emerged as a promising solution, yet current designs limit their performance. Existing single-agent systems scale poorly to long-horizon, multi-step, and realistic tasks, whereas multi-agent systems typically execute predefined workflows that cannot capture the iterative, non-linear nature of a real drug discovery campaign. Here, we present MADDIE, a multi-agent drug discovery engine. It comprises a central agent, Orchestrator, and four specialized worker agents: HitIdentifier, LeadOptimizer, Synthesizer, and Evaluator, each representing a critical step in a drug discovery campaign. Under Orchestrator's orchestration and coordination, these worker agents are composed into flexible workflows, enabling MADDIE to adapt its strategy and navigate the drug discovery pipeline end-to-end. Across 50 protein targets, MADDIE produces molecules satisfying key property requirements and their synthesis routes, outperforming existing agentic systems, generative models, and general-purpose LLMs by a wide margin. Our analysis reveals how these agents collaborate and that MADDIE can employ different strategies across targets to achieve the campaign's goal, while still being efficient in time and tokens. Together, they demonstrate MADDIE as a low-cost, effective, and efficient autonomous engine for accelerating drug discovery. Code and dataset is available via https://github.com/ninglab/MADDIE.

Abstract

Explore similar work

MADDIE: Multi-Agent Drug Discovery Engine

Multi-Agent Orchestration for High-Throughput Materials Screening on a Leadership-Class System

MDForge: Agentic Molecular Dynamics Pipeline Design under Sparse Simulator Feedback