Specifying Update Policies for Nonmonotonic Knowledge Bases

Recently, several approaches for updating knowledge bases represented as logic programs have been proposed. We present a generic framework for declarative specifications of update policies, which is built upon such approaches. It extends the LUPS language for update specifications, and incorporates the notion of events into the framework. An update policy allows an agent to flexibly react upon new information, arriving as an event, and perform suitable changes of its knowledge base. The framework compiles update policies to logic programs by means of generic translations, and can be instantiated in terms of different concrete update approaches. It thus provides a flexible tool for designing adaptive reasoning agents.