Belief ascription under bounded resources (Q987466)

This is a study in multi-agent epistemic temporal logic with bounded resources. It is observed that when the agents' computational powers and memory are bounded, then they may not be able to ascribe beliefs to each other correctly even `in the limit', and even if they know about each other's inference engines (e.g. that they can perform resolution). The paper shows, however, that if agents also know about each other's `reasoning strategies', understood as a preference for deriving certain formulae over others, then they can correctly ascribe beliefs in the limit. A final section usefully shows where the contribution fits in to the existing literature.