 | Wooldridge treats multiagent systems as environments where autonomous programs share worlds — coordination, negotiation, and communication protocols decide whether collective intelligence compounds or collides. Wooldridge First published 2002 - 2nd ed. (2009) · Wiley · ISBN 9780470519460 |
|---|
Key principles
- Autonomous agent model — Perceive, decide, act loops with local goals — global outcomes emerge from interaction, not central script.
- Coordination problem — Shared tasks need role division and timing — failure shows up as duplicate work or deadlock, not bad intentions.
- Communication acts — Messages carry commitments — protocol design matters as much as individual agent IQ.
- Negotiation and allocation — Scarce resources require explicit bargaining — game structure shapes fair and stable agreements.
- Environment state — World model is shared or partially observable — hand-offs fail when agents inherit stale context.
- Engineering trade-offs — Centralized control vs decentralised resilience — pick architecture to match failure tolerance.
Core science
Wooldridge’s textbook is the standard multiagent systems spine — extending Russell and Norvig’s single-agent frame (Artificial Intelligence) with distributed AI, coordination languages, and mechanism design. On the Codex shelf it pairs with Agent orchestration for 2026 workflow wiring — context contracts, role boundaries, and explicit hand-offs between capture, synthesis, and execution agents.
Application
Reach for this volume when multi-agent coordination or AI orchestration needs a textbook spine — before stacking tools without shared state. Open Agent orchestration on the Atlas shelf for the mechanism map.