Improvements on graph path queries: expression, evaluation, and minimum-weight satisfiability

Abstract

We deal with three problems related to graph path queries. Most current graph query languages support regular path queries. However, some applications such as source-code analysis and genetics require context-free path queries. Context-free path queries use context-free languages. There is no standard notation for context-free languages simpler than context-free grammars. The evaluation of a context-free path query is more complex than a regular path query. Moreover, in some applications, it is desired to have the minimum graph that preserves answers to a given path query. To address each of those problems, we: (1) develop an alternative notation for expressing context-free languages; (2) design, implement and experiment with a context-free path query evaluation algorithm; and (3) formalize the formal-language-constrained graph minimization problem, for which we design solutions for the cases where the formal language is regular or context-free.