580 California St., Suite 400
San Francisco, CA, 94104
Key research themes
1. How can cooperation strategies be structured and controlled in multi-agent systems for effective coordination and task execution?
This theme explores the design of agent architectures and control models to manage cooperation complexity in multi-agent systems (MAS), especially focusing on hierarchical, modular, or multi-resolution frameworks that enable coordination, negotiation, and task distribution among agents to achieve collective goals efficiently.
Key finding: Introduced the MRCC model which decomposes cooperative control into hierarchical levels—system, micro-social, and individual agent control layers—enabling flexible management of cooperation complexity. The internal agent... Read more
Key finding: Provided a comprehensive survey clarifying that software agents in MAS operate autonomously and proactively, often with roles and agendas that necessitate coordination for complex real-world problem solving. Highlighted the... Read more
Key finding: Proposed the High-level Multi Agent Petri Net (HMAP) as a formal framework to model, analyze, and simulate MAS behavioral dynamics including roles, collaborations, and asynchronous interactions. Demonstrated through formal... Read more
Key finding: Developed a token-based decentralized coordination method (CABS) that enables agents to dynamically regulate job requests and execution to avoid starvation of bottleneck resources in large-scale systems. This approach... Read more
Key finding: Presented protocols for cooperative agents based on distributed constraint satisfaction problems (CSP) and mechanism design for strategic agent interaction. Introduced algorithms (e.g., asynchronous backtracking and... Read more
2. What optimization frameworks and game-theoretic models underpin cooperative and competitive behaviors in multi-agent systems?
This theme investigates mathematical and algorithmic frameworks to model optimization and strategic decision-making in MAS, particularly focusing on distributed global optimization, cooperative and non-cooperative games, and their impact on agent interactions—addressing both collaborative and competitive scenarios with applications in privacy protection, equilibrium computation, and large-scale MAS behavior analysis.
Key finding: Proposed the MANGO environment where different optimization algorithms are embedded within autonomous agents that cooperate via adaptive communication protocols rather than fixed interaction patterns. This autonomous... Read more
Key finding: Comprehensively surveyed distributed optimization techniques (including online and federated optimization) and game-theoretic models (static/dynamic and cooperative/non-cooperative games) to characterize various MAS... Read more
Key finding: Developed a novel communication management framework based on adaptive sensing ranges for agents, combined with predictive control to maintain network connectivity while saving communication energy. This approach demonstrates... Read more
Key finding: Advanced the use of colored Petri nets for formal representation of interaction protocols, enabling agents to dynamically maintain and adapt conversation states. Addressed the challenge of protocol learning and modification... Read more
Key finding: Applied cooperative MAS to strategic scanning processes, integrating actors' interactions for incremental knowledge building. Emphasized multi-agent distributed optimization of information quality, validation, and security in... Read more
3. How can multi-agent systems be leveraged in applications to achieve cooperative decision-making and control in dynamic, distributed environments?
This theme focuses on practical MAS applications in robotics, distributed control systems, and real-world infrastructure management, emphasizing architectures and methodologies that enable coordination, adaptive role allocation, and cooperative decision-making under dynamic conditions, highlighting the use of MAS for problem-solving in navigation, control, and resource management.
Key finding: Implemented a distributed multi-agent control system for municipal water distribution, demonstrating adaptability to uncertain demand and operational perturbations. Showed that distributed agent-based pump scheduling can... Read more
Key finding: Reviewed MAS applications in mobile robotics, identifying cooperative navigation, path planning, obstacle avoidance, and negotiation protocols as core capabilities achieved through MAS frameworks. Emphasized the importance of... Read more
Key finding: Developed a fuzzy logic-based decision making framework enabling cooperative soccer robots to select roles and actions dynamically. Demonstrated that hierarchical decision-making based on robot situational awareness and team... Read more
Key finding: Provided fundamental MAS concepts emphasizing agent autonomy, social behavior, proactiveness, and situatedness. Highlighted how these attributes underpin distributed decision-making and control in open, uncertain... Read more
Key finding: Surveyed control algorithms for achieving consensus and synchronization in networked MAS with communication constraints. Highlighted how communication topology and local interaction protocols influence emergent coordinated... Read more