Selected work

Representative Papers

A Grid-Aware Agent-Based Model for Analyzing Electric Vehicle Charging Systems

Apr 30, 2026

Existing simulation models struggle to simultaneously capture user charging behavior, charging infrastructure constraints, and grid interactions, limiting systematic evaluation of electric vehicle (EV) charging systems. This work proposes a grid-aware multi-agent discrete-event simulation framework that, for the first time, couples user-level charging dynamics with facility-level grid responses and introduces a shared energy sandbox mechanism to enable aggregated power regulation. The model integrates heterogeneous EV behaviors, charger constraints, and real-time power allocation algorithms, supporting flexible configuration of infrastructure and scheduling strategies. Experiments in a representative workplace scenario demonstrate that the combination of charging strategies and charger types significantly affects service performance, facility utilization, and grid load, underscoring the high degree of scenario dependency in infrastructure design and deployment.

0 citationsRead paper

Flocking Behavior: An Innovative Inspiration for the Optimization of Production Plants

Aug 27, 2025

Frequent machine-type switching in semiconductor manufacturing introduces significant scheduling complexity and prolonged setup delays. Method: This paper proposes a distributed optimization approach based on swarm intelligence, pioneering the application of the bio-inspired “Boids” flocking algorithm to production scheduling. Leveraging decentralized, bottom-up local sensing and simple heuristic rules, the method enables self-organized coordination across hybrid environments comprising both single-batch and batch-processing machines—without requiring global information or centralized computation. It dynamically regulates job flows to avoid conflicts and adapt to real-time shop-floor disturbances. Contribution/Results: Experimental results demonstrate substantial improvements in system flexibility and responsiveness. The approach exhibits excellent scalability and runtime stability in large-scale scenarios, offering a novel paradigm for real-time, adaptive scheduling in complex discrete manufacturing systems.

0 citationsRead paper

LLM-Powered Swarms: A New Frontier or a Conceptual Stretch?

Jun 17, 2025

This paper investigates the paradigm shift induced by deploying large language models (LLMs) as autonomous agents in swarm systems, contrasting their properties against classical swarm intelligence (e.g., Boids, ant colony optimization) in terms of decentralization, scalability, and emergence. Method: We formally define the conceptual boundaries of LLM swarms, highlighting fundamental distinctions in computational overhead, coordination mechanisms, and semantic emergence logic. Integrating classical swarm algorithms with multi-scale LLM deployment (cloud-edge-device), we propose a three-dimensional evaluation framework assessing behavioral accuracy, latency, and resource consumption. Contribution/Results: Empirical analysis reveals that LLM swarms exhibit strong semantic reasoning and abstract collaborative capabilities but incur substantially higher latency and computational cost. Consequently, they are best suited for high-semantic, low-real-time applications. Based on these findings, we introduce the first taxonomy for LLM swarm applicability, advancing a theoretical redefinition of “swarm” in the AI era.

0 citationsRead paper

Recent publications

Latest Papers

A Grid-Aware Agent-Based Model for Analyzing Electric Vehicle Charging Systems

Apr 30, 2026

Existing simulation models struggle to simultaneously capture user charging behavior, charging infrastructure constraints, and grid interactions, limiting systematic evaluation of electric vehicle (EV) charging systems. This work proposes a grid-aware multi-agent discrete-event simulation framework that, for the first time, couples user-level charging dynamics with facility-level grid responses and introduces a shared energy sandbox mechanism to enable aggregated power regulation. The model integrates heterogeneous EV behaviors, charger constraints, and real-time power allocation algorithms, supporting flexible configuration of infrastructure and scheduling strategies. Experiments in a representative workplace scenario demonstrate that the combination of charging strategies and charger types significantly affects service performance, facility utilization, and grid load, underscoring the high degree of scenario dependency in infrastructure design and deployment.

0 citationsRead paper

Flocking Behavior: An Innovative Inspiration for the Optimization of Production Plants

Aug 27, 2025

Frequent machine-type switching in semiconductor manufacturing introduces significant scheduling complexity and prolonged setup delays. Method: This paper proposes a distributed optimization approach based on swarm intelligence, pioneering the application of the bio-inspired “Boids” flocking algorithm to production scheduling. Leveraging decentralized, bottom-up local sensing and simple heuristic rules, the method enables self-organized coordination across hybrid environments comprising both single-batch and batch-processing machines—without requiring global information or centralized computation. It dynamically regulates job flows to avoid conflicts and adapt to real-time shop-floor disturbances. Contribution/Results: Experimental results demonstrate substantial improvements in system flexibility and responsiveness. The approach exhibits excellent scalability and runtime stability in large-scale scenarios, offering a novel paradigm for real-time, adaptive scheduling in complex discrete manufacturing systems.

0 citationsRead paper

LLM-Powered Swarms: A New Frontier or a Conceptual Stretch?

Jun 17, 2025

This paper investigates the paradigm shift induced by deploying large language models (LLMs) as autonomous agents in swarm systems, contrasting their properties against classical swarm intelligence (e.g., Boids, ant colony optimization) in terms of decentralization, scalability, and emergence. Method: We formally define the conceptual boundaries of LLM swarms, highlighting fundamental distinctions in computational overhead, coordination mechanisms, and semantic emergence logic. Integrating classical swarm algorithms with multi-scale LLM deployment (cloud-edge-device), we propose a three-dimensional evaluation framework assessing behavioral accuracy, latency, and resource consumption. Contribution/Results: Empirical analysis reveals that LLM swarms exhibit strong semantic reasoning and abstract collaborative capabilities but incur substantially higher latency and computational cost. Consequently, they are best suited for high-semantic, low-real-time applications. Based on these findings, we introduce the first taxonomy for LLM swarm applicability, advancing a theoretical redefinition of “swarm” in the AI era.

0 citationsRead paper