A retrieval data structure implements a mapping from a set S of n keys to range R={0,1} r , e.g. given by a list of key-value pairs (x,v) S×R, but an element outside S may be mapped to any value. Asymptotically, minimal perfect hashing allows to build such a data structure that needs nlog2 e+nr+o(n) bits of memory and has constant evaluation time. Recently, data structures based on other approaches have been proposed that have linear construction time, constant evaluation time and space consumption O(nr) bits or even (1+ε)nr bits for arbitrary ε>0. This paper explores the practicability of one such theoretically very good proposal, bridging a gap between theory and real data structures. © 2009 Springer Berlin Heidelberg.
CITATION STYLE
Aumüller, M., Dietzfelbinger, M., & Rink, M. (2009). Experimental variations of a theoretically good retrieval data structure. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5757 LNCS, pp. 742–751). https://doi.org/10.1007/978-3-642-04128-0_66
Mendeley helps you to discover research relevant for your work.