The field of cyber-physical systems and social simulations is rapidly evolving, with a focus on developing more flexible, scalable, and intelligent solutions. Researchers are exploring the use of knowledge graphs, multi-agent systems, and large language models to improve the integration of physical and digital environments, simulate complex human behaviors, and generate realistic social scenarios. Noteworthy papers in this area include KG-MAS, which introduces a knowledge graph-enhanced multi-agent infrastructure for coupling physical and digital robotic environments, and SocioBench, which presents a comprehensive benchmark for evaluating the alignment of large language models with real-world social attitudes. Other notable works, such as StoryBox and Evolution in Simulation, demonstrate the potential of multi-agent simulations and large language models for generating dynamic, immersive stories and simulating complex educational dynamics. Overall, the field is moving towards more open-ended, adaptive, and socially aware simulations that can capture the complexity of real-world systems.
Advancements in Cyber-Physical Systems and Social Simulations
Sources
KG-MAS: Knowledge Graph-Enhanced Multi-Agent Infrastructure for coupling physical and digital robotic environments
FIDRS: A Novel Framework for Integrated Distributed Reliable Systems
SocioBench: Modeling Human Behavior in Sociological Surveys with Large Language Models
Evolution in Simulation: AI-Agent School with Dual Memory for High-Fidelity Educational Dynamics
Valid Survey Simulations with Limited Human Data: The Roles of Prompting, Fine-Tuning, and Rectification
Survey Response Generation: Generating Closed-Ended Survey Responses In-Silico with Large Language Models
StoryBox: Collaborative Multi-Agent Simulation for Hybrid Bottom-Up Long-Form Story Generation Using Large Language Models
Runtime Composition in Dynamic System of Systems: A Systematic Review of Challenges, Solutions, Tools, and Evaluation Methods
Scrutiny new framework in integrated distributed reliable systems
Too Open for Opinion? Embracing Open-Endedness in Large Language Models for Social Simulation
Static Sandboxes Are Inadequate: Modeling Societal Complexity Requires Open-Ended Co-Evolution in LLM-Based Multi-Agent Simulations