Belief-Desire-Intention (BDI)

Terms

Beliefs: facts that an agent believes about the world, like knowledge. But it is not necessary to be always true.

Desires: are goals and preferred end states. An agent may have many desires, possibly in conflict.

Intentions: are the agent’s commitments to its desire. They shall be consistent and without conflict.

Rational Agents:

A rational agent is an agent that has clear preferences, models uncertainty via expected values of variables or functions of variables, and always chooses to perform the action with the optimal expected outcome for itself from among all feasible actions. A rational agent can be anything that makes decisions, typically a person, firm, machine, or software. (Wikipedia)

Some statements

Intentions are fundamentally based on the notion of commitments

Commitments relate to satisficing (good enough), and optimal to the best solution

Intentions are meant to stabilize decision making (Bratman)

BDI logics

• Cohen and Levesque logic
• Rao and Georgeff’s belief-desire-intention logic

BDI Architecture

• Belief store
• Goal store
• Plan-library
• Intention hierarchy

Applications

• BDI agent models for test generation (Prof. Kerstin Eder)
  • BDI agents vs. automata for model-based test generation (see “Model-based Test Generation for Robotic Software:Automata versus Belief-Desire-Intention Agents”)

References:

• M. E. Bratman. Intention, plans, and practical reason. Harvard University Press, Cambridge Massachusetts, 1987
• “Modeling rational agents within a BDI architecture” by Rao and Georgeff
• “BDI Agents: From Theory to Practice” by Rao and Georgeff
• “The Belief-Desire-Intention Model of Agency” by Michael Georgeff et. al
• “CHAPTER 2: Beliefs, Desires, Intentions (BDI Logic)” by Milind Tambe
• P. Cohen and H. Levesque. Intention is choice with commitment. Artificial Intelligence, 42(2-3):213–261, 1990
• “AgentSpeak(L): BDI Agents speak out in a logicalcomputable language” by Rao regarding BDI architectures