Evolving graph-structures and their implicit computational complexity
Résumé
Dynamic data-structures are ubiquitous in programming, and they use extensively underlying directed multi-graph structures, such as labeled trees, DAGs, and objects. This paper adapts well-established static analysis methods, namely data ramification and language-based information flow security, to programs over such graph structures. Our programs support the creation, deletion, and updates of both vertices and edges, and are related to pointer machines. The main result states that a function over graph-structures is computable in polynomial time iff it is computed by a terminating program whose graph manipulation is ramified, provided all edges that are both created and read in a loop have the same label.
Domaines
Complexité [cs.CC]Origine | Fichiers produits par l'(les) auteur(s) |
---|
Loading...