GDC is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

View, browse and sort the ever-growing list of sessions by day, pass type, topic, and format. With this Session Viewer, you can view GDC 2023 session details and speakers, and share your favorites via social media. You will be able to build your schedule and access it during the show via export or Mobile App, once live. Sessions do fill up and seating is first come, first serve, so arrive early to sessions that you would like to attend.

AI Summit: Let Your Agents Plan Together: Multi-Agent Cooperation With GOAP

Gautier Boeda  (AI Engineer, SQUARE ENIX CO., LTD.)

Location: Room 2002, West Hall

Date: Monday, March 20

Time: 3:50 pm - 4:20 pm

Pass Type: All Access Pass, Summits Pass

Topic: Programming

Format: Lecture

Vault Recording: Video

Audience Level: Intermediate

In this session, Gautier Boeda, AI Engineer at Square Enix, covers a multi-agent cooperation system in a world where there is no super-entity that manages the agents, and no shared memory between all agents.

This makes the cooperation between AI-agents more natural and believable than a classic cooperation system—based on a super entity managing all agents with a shared memory. The agents can only plan from what they know, have to do hypotheses and talk to other agents to get help. The developers do not need to create fake conversations or behaviors in order to make the cooperation believable, as if the agent thought about it by itself. This system is based on GOAP planning technology with a messaging system.

Examples will be taken from two different experimental projects: a character interaction project and a RPG battle project.

Takeaway

Attendees will learn how to design a multi-agent cooperation system when there is no super-entity to manage the agents and think for them, or any shared memory to make the agents' communications and behaviors more believable and natural in order to enhance immersion for the player.

Intended Audience

This is for anyone curious about multi-agent systems, especially in the context where there is no super-entity managing the agents. In addition, anyone interested about communication between agents, negotiations, or planning with perception-limited knowledge will benefit from this talk.