Cumulative Space in Black-White Pebbling and Resolution

QC 20170509
We study space complexity and time-space trade-offs with a focus not on peak memory usage but on overall memory consumption throughout the computation. Such a cumulative space measure was introduced for the computational model of parallel black pebbling by [Alwen and Serbinenko 2015] as a tool for obtaining results in cryptography. We consider instead the nondeterministic black-white pebble game and prove optimal cumulative space lower bounds and trade-offs, where in order to minimize pebbling time the space has to remain large during a significant fraction of the pebbling. We also initiate the study of cumulative space in proof complexity, an area where other space complexity measures have been extensively studied during the last 10-15 years. Using and extending the connection between proof complexity and pebble games in [Ben-Sasson and Nordström 2008, 2011], we obtain several strong cumulative space results for (even parallel versions of) the resolution proof system, and outline some possible future directions of study of this, in our opinion, natural and interesting space measure.

de Rezende, Susanna F.KTH,Teoretisk datalogi, TCS(Swepub:kth)u15llzty (författare)
Nordström, Jakob,1972-KTH,Teoretisk datalogi, TCS(Swepub:kth)u1mbi3f0 (författare)
Vinyals, MarcKTH,Teoretisk datalogi, TCS(Swepub:kth)u1rb5z2h (författare)
KTHTeoretisk datalogi, TCS (creator_code:org_t)

Ingår i:Leibniz International Proceedings in Informatics, LIPIcs: Schloss Dagstuhl- Leibniz-Zentrum fur Informatik GmbH, Dagstuhl Publishing9783959770293

Av författaren/redakt...: Alwen, Joël; de Rezende, Susa ...; Nordström, Jakob ...; Vinyals, Marc

Om ämnet

Kungliga biblioteket hanterar dina personuppgifter i enlighet med EU:s dataskyddsförordning (2018), GDPR. Läs mer om hur det funkar här.
Så här hanterar KB dina uppgifter vid användning av denna tjänst.