Set differentiation: a method for the automatic generation of filtering algorithms
Abstract
We present a method and its implementation in the GAP system for the automatic generation of filtering algorithms. The evaluation of left-hand sides of rules relies on the matching of condition elements with working memory elements. The filtering is the inference engine phase that performs this matching. Our automatic generation method is based on set differentiation, taking into account both qualitative and quantitative aspects. We present GAP's architecture and show how generic filtering algorithm skeletons are built using set differentials.