Modeling emotions for choosing between deliberation and action

Agents that operate in the real world have to make decisions on how long they can deliberate before they act. If the agent deliberates for too long, the agent may miss a deadline or the environment may change such that the preconditions for acting may no longer hold true. If the agent acts too quickly without proper deliberation, it may miss opportunities or may even perform the wrong action. Artificial agents with a metacognitive ability to monitor and influence its deliberation and action can potentially make intelligent decisions on when to stop deliberating and start acting. This research is motivated by the evidence that suggests the role of emotions in the cognitive activities of humans. This paper describes how a model of emotions can be applied to enhance the metacognitive capabilities of an artificial agent to choose between deliberation and action. The emotional model is illustrated within the domain of an air traffic control simulation system.