Although expectations play an important role in designing cognitive agents, monitoring for agent expectations is not explicitly being handled in most common agent programming environments. There are techniques for monitoring fulfilment and violation of agent expectations, however they are not linked with common agent programming environments so that agents can be easily programmed to respond to these circumstances. This paper investigates how to delegate this aspect of agent practical reasoning to an expectation monitoring tool integrated with a BDI agent platform. We exemplify this using the Jason BDI agent interpreter by extending it with built-in actions to initiate and terminate monitoring of expectations. This delegation enables agents to monitor for the fulfilment and violation of their expectations without relying on a centralised monitoring mechanism. This way, it is possible for agents to have plans that respond to the identified fulfilments and violations of their expectations.
展开▼
