Submission Deadline: 31 March 2027 View: 55 Submit to Special Issue
Prof. Wenbing Zhao
Email: w.zhao1@csuohio.edu
Affiliation: Department of Electrical and Computer Engineering, Cleveland State University, Cleveland, United States
Research Interests: machine learning, LLM, LLM agent, sports analytics
Prof. Pan Wang
Email: wangpan@njupt.edu.cn
Affiliation: School of Modern Posts, Nanjing University of Posts and Telecommunications, Nanjing, China
Research Interests: AI enabled network management, information and network security, Internet of Things, AI for recommender system, future communications and networks, smart grid
The evolution of computational science is transitioning from human-scripted simulations to autonomous, AI-driven workflows. While traditional data-driven surrogates like Physics-Informed Neural Networks (PINNs) have improved prediction speeds, the emergence of Large Language Models (LLMs) and foundation models introduces a cognitive layer previously missing in engineering. These architectures are no longer just "calculators" but "orchestrators" capable of reasoning through physical constraints, generating complex simulation codes, and steering digital twins in real-time.
This special issue focuses on the cognitive integration of AI into computational modeling. We seek pioneering research where LLMs act as the primary interface or autonomous agent for modeling, simulation, and scientific discovery. This includes innovative approaches that leverage the reasoning capabilities of foundation models, utilizing agents to fine-tune and adapt these large-scale architectures for specialized, physics-heavy simulation tasks. The goal is to move beyond general machine learning applications and explore how agentic AI can redefine the "syntax" of engineering design and physical system optimization.
Topics of Interest
We invite original research and review articles that demonstrate a clear "Human-in-the-loop" or "Agentic" focus within computational engineering, including:
· Autonomous Simulation Agents: LLM-driven frameworks for the automated synthesis, debugging, and verification of numerical codes (e.g., FEM, CFD, and MD).
· Agent-Led Adaptation of Foundation Models: Methods for using autonomous agents to fine-tune, prune, or adapt large-scale foundation models for specific engineering domains and physical laws.
· Cognitive Digital Twins: Foundation model-powered conversational agents for real-time monitoring, multi-modal sensor steering, and interactive control of physical systems.
· LLM-Guided Scientific Discovery: Agentic workflows that use automated reasoning to hypothesize, test, and refine material properties or structural designs.
· Natural Language Interfaces for Simulation: Developing sophisticated interfaces that translate complex engineering intent into executable physical models and optimization protocols.
· Multi-Modal Physics Orchestration: Integrating text-based design requirements with 3D CAD geometries and sensor feedback through unified foundation model architectures.
· Agent Verification & Safety: Methods for ensuring the physical consistency and safety of AI-generated engineering decisions and simulation outputs.
Login
Register
Submit a Paper
Propose a Special lssue