We investigate the complexity of obtaining sparse descriptions for sets in various reduction classes to sparse sets. Let A be a set in a certain reduction class R,.(SPARSE). Then we are interested in finding upper bounds for the complexity (relative to A) of sparse sets S such that A ε Rr(S). By establishing such upper bounds we are able to derive the lowness of A. In particular, we show that if a set A is in the class Rphd(Rpc(SPARSE)) then A is in Rphd(Rpc(S)) for a sparse set S ε NP(A). As a consequence we can locate Rphd(Rpc(SPARSE)) in the EL⊖3 level of the extended low hierarchy. Since Rphd(Rpc(SPARSE)) ⊇ Rphd(Rpc(SPARSE)) this solves the open problem of locating the closure of sparse sets under bounded truth-table reductions optimally in the extended low hierarchy. Furthermore, we show that for every A ε Rpd(SPARSE) there exists a sparse set S ε NP(A ⊕ SAT)/F⊖p2(A) such Lhat A ε Rpd(S). Based on this we show that Rp1-tt(Rpd(SPARSE)) is in EL⊖3 Finally, we construct for every set A ε Rpc(TALLY) ∩Rpd(TALLY) (or equivalently, A ε IC[log, poly], as shown in [AHH+92]) a tally set T ε P(A ⊕ SAT) such that A 9 Rpc(T) ∩ Rpd(T). This implies that the class IC[log, poly] of sets with low instance complexity is contained in EL∑1.
CITATION STYLE
Arvind, V., Köbler, J., & Mundhenk, M. (1992). Lowness and the complexity of sparse and tally descriptions. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 650 LNCS, pp. 249–258). Springer Verlag. https://doi.org/10.1007/3-540-56279-6_78
Mendeley helps you to discover research relevant for your work.