FIT3226 · Agentic and distributed AI: Multi-agent systems, swarms and artificial life

Overview

This unit explores artificial intelligence arising from interactions between autonomous intelligent agents. You will examine three perspectives: agentic AI, where agents are designed top-down with defined roles yet make independent decisions; multi-agent systems, where agents coordinate, negotiate, and sometimes compete while planning and acting toward goals; and emergent and artificial life approaches, where simple, locally specified interactions produce complex, often unpredictable global behaviour. Grounded in systems-thinking and complex adaptive systems, the unit highlights decentralisation, feedback, adaptation, and unintended outcomes. You will design and develop agentic and multi-agent AI systems capable of reasoning, planning, and multi-step action in dynamic environments, while analysing emergent behaviour, conflict, robustness, and the implications of distributed intelligence in AI systems.

Offerings

The Handbook publishes no offerings for this unit.

Assessment

The Handbook publishes no assessment items for this unit yet.

Assessment details may change. Please refer to the assessment information in Moodle closer to the start of the teaching period.

Requisites

Learning outcomes

1. Configure computational environments, frameworks and simulation tools to build and evaluate agentic and multi‑agent AI systems operating in dynamic conditions;
2. Apply systems-thinking methods to analyse and compare agentic AI, coordinated multi-agent systems, and emergence/artificial-life approaches, explaining how local rules and interactions produce global behaviours in complex adaptive systems;
3. Design system architectures, coordination protocols and rule‑based interaction models that balance decentralisation, adaptability and control in distributed AI systems;
4. Implement interacting agents that can plan, reason, and execute multi-step actions (including tool use where appropriate), integrating coordination logic and handling partial information or dynamic environments;
5. Engineer, run, and interpret evaluations of multi-agent behaviour (e.g., simulations, ablations, and performance/emergence metrics) to iteratively improve robustness, coordination quality, and goal attainment.

Workload

No workload detail published.

Student experience, not official rules. Nothing here changes what the Handbook says.