An architecture for intelligent control based on epistemic utility theory

Epistemic utility theory is used to define an architectural model of intelligent control. The model provides each agent with an epistemic system, to account for its state of knowledge, its goals and values, its beliefs, its willingness to risk error, and the existence of incomplete or contradictory evidence. Agents make decisions on a local, rather than a global basis, thus preserving functionality without requiring optimality. Multiple agents may share joint epistemic systems, making possible the formulation of cooperative, contradictory, or mixed decision and control strategies.