Updating Contexts

Different formalisms for contexts have been successfully used to represent and reason about distributed knowledge. In these formalisms, knowledge is represented as a set of contexts, each representing a piece of the whole knowledge. Contexts have been proved particularly adapt to deal with heterogeneous distributed knowledge, but, in order to deal with the change of such a knowledge, they must be extended. In particular, they must cope with the problems of adding, deleting, or changing facts in a context, and computing the effect of this change in the other contexts. The current approaches to belief revision and multi-agent belief revision can be helpful, but they do not provide a satisfactory treatment of heterogeneity. We provide a formal definition of the operation of updating a context, and we define an algorithm that computes the effects of updating a context on other related contexts. We take a semantic perspective, i.e., each context is formalized as a set of possible partial models of the world