Strategies for repeated games with subsystem takeovers: implementable by deterministic and self-stabilizing automata (extended abstract)

Shlomi Dolev, Elad Michael Schiller, Paul G. Spirakis, Philippas Tsigas

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Systems of selfish-computers, such as the Internet, are subject to transient faults due to hardware/software temporal malfunctions; just as the society is subjected to human mistakes due to a moment of weakness. Game theory uses punishment for deterring improper behavior. Due to faults, selfish-computers may punish well-behaved ones. This is one of the key motivations for forgiveness that follows any effective and credible punishment. Therefore, unplanned punishments must be proven to have ceased in order to avoid infinite cycles of unsynchronized behavior of “tit for tat”. We investigate another aspect of selfish-computer systems. We consider the possibility of subsystem takeover, say, by the use of hostile malware. The takeover may lead to joint deviations coordinated by an arbitrary selfish-computer that controls an unknown group of subordinate computers. We present strategies that deter the coordinator (and its subordinates) from deviating in infinitely repeated games. We construct deterministic and finite automata that implement these strategies with optimal complexity. Moreover, we prove that all unplanned punishments eventually cease by showing that the automata can recover from transient faults.
Original languageEnglish
Title of host publicationProceedings of the 2nd International Conference on Autonomic Computing and Communication Systems, Autonomics 2008, September 23-25, 2008, Turin, Italy
EditorsAntonio Manzalini
PublisherInstitute for Color Science and Technology (ICST)
Pages37
Number of pages1
DOIs
StatePublished - 2008

Fingerprint

Dive into the research topics of 'Strategies for repeated games with subsystem takeovers: implementable by deterministic and self-stabilizing automata (extended abstract)'. Together they form a unique fingerprint.

Cite this