Recursive Plans for Information Gathering

Generating query-answering plans for information gathering agents requires to translate a user query, formulated in terms of a set of virtual relations, to a query that uses relations that are actually stored in information sources. Previous solutions to the translation problem produced sets of conjunctive plans, and were therefore limited in their ability to handle information sources with binding-pattern limitations, and to exploit functional dependencies in the domainmodel. As a result, these plans were incomplete w.r.t. sources encountered in practice (i.e., produced only a subset of the possible answers). We describe the novel class of recursive information gathering plans, which enables us to settle two open problems. First, we describe an algorithm for nding a query plan that produces the maximal set of answers from the sources in the presence of functional dependencies. Second, we describe an analogous algorithm in the presence of binding-pattern restrictions in the sources, which was not possible without recursive plans.