Rule management in expert database systems

Expert database systems combine database and expert systems technologies to support the effective management of both rules and data. This paper studies rule processing strategies in expert database systems involving rules that are conditional on joins of relational data. Auxiliary constructs for processing join rules are proposed, and a framework of join rule processing strategies is developed. Cost functions of several strategies are derived based on a stochastic model that characterizes the arrival processes of transactions and queries to the database. Performance evaluation shows that the proposed data constructs and strategies provide an effective method for processing rules.