Multi-Agent Systems & Coordination in Python Training Course

Introduction to Multi-Agent Systems

• Overview of agents, environments, and interaction models
• Cooperation, competition, and autonomy in agentic systems
• Applications in logistics, robotics, and decision-making

Core Concepts of Agent Architecture

• Reactive vs. deliberative agents
• Communication protocols and coordination models
• Knowledge representation and shared state

Implementing Agents in Python

• Building agents using the Mesa framework
• Modelling environments and interactions
• Simulating agent behaviour and visualisation

Coordination and Communication

• Message passing and shared memory architectures
• Negotiation, consensus, and task allocation
• Coordination algorithms (contract net, market-based, swarm models)

Learning and Adaptation in Multi-Agent Systems

• Reinforcement learning for multiple agents
• Cooperative vs. competitive learning dynamics
• Utilising PettingZoo and Stable-Baselines3 for MARL

Distributed Computing and Scaling

• Using Ray for distributed multi-agent simulations
• Managing concurrency and synchronisation
• Parallelising computation and handling shared resources

Human–Agent Collaboration

• Designing interfaces for human-in-the-loop coordination
• Hybrid workflows with AI-assisted decision support
• Ethical and operational considerations

Capstone Project

• Design and implement a multi-agent system in Python
• Demonstrate coordination and learning among agents
• Present simulation results and performance insights

Summary and Next Steps