Course contentsShow
← AI Engineering
Lesson 697 of 1,88617. Multi-Agent SystemsFree lesson

Blackboard Architecture for Shared State

Use a shared data structure where agents read and write information to coordinate asynchronously.

What You'll Learn

You'll learn how a blackboard architecture lets multiple agents coordinate by reading and writing to a shared data structure—without needing to talk directly to each other.


The Core Idea

Imagine a classroom where students (agents) all work on solving parts of a big problem. Instead of constantly asking each other questions, they write their findings on a shared blackboard. Anyone can read what others wrote, add new insights, or update existing information. No one "owns" the conversation—they all contribute asynchronously to the shared knowledge space.

In multi-agent systems, a blackboard is a centralized data structure (like a dictionary, database, or object store) where agents post observations, partial results, hypotheses, or task updates. Other agents monitor the blackboard, react to changes, and contribute their own work. This decouples agents from each other—they only need to know how to read/write the blackboard, not who else is using it.

Why Use a Blackboard?

  • Asynchronous coordination: Agents don't block waiting for replies
  • Loose coupling: Agents don't need to know each other's identities
  • Incremental problem-solving: Each agent adds a piece; the solution emerges collectively
  • Flexibility: Easy to add or remove agents without changing communication logic

When It Fits

Blackboards work well when:

  • Multiple agents solve different parts of a complex problem (e.g., data analysis, research, planning)
  • Agents have complementary skills and need to share partial results
  • You want agents to "pick up work" opportunistically rather than follow strict handoffs

Key Takeaway: Blackboard architecture uses a shared data structure as a coordination hub, letting agents contribute asynchronously without direct messaging—ideal for collaborative, incremental problem-solving.