Abstract

: Situations in which the information about a given domain is partial are common in many AI applications. In planning and analysis of scenarios involving partial information, the state of knowledge of an intelligent agent in such circumstances becomes important. This paper addresses the problem of characterizing this state of knowledge, with the emphasis on the single-agent case. We give a number of equivalent ways to characterize this state of knowledge, as well as an algorithm for computing the formulas that are true in this state. The relationship between this work and related works by Stark, Konolige, and Moore is discussed. 1. Introduction This research was originally motivated by the question of how communication in a distributed system changes the state of knowledge of the processors in the system (cf. [HM1]). Answering such a question clearly requires characterizing a processor's state of knowledge at a given point in time. To see some of the difficulties here, consider a pr...

Citations