05375029

j

05375029 Grebla, Horea Adrian Game theory in data allocation for distributed databases. Acta Univ. Apulensis, Math. Inform. 11, 325-331 (2006). 2006 "1 Decembrie 1918" University of Alba Iulia, Department of Mathematics and Informatics, Alba Iulia EN distributed databases balancing games Nash equilibrium Summary: This paper studies distributed database performance optimization by mapping data fragmentation and allocation to load balancing games. The system is modelled as a set of network locations (agents) who share several resources (data fragments). The aim of the agents is to follow a strategy that maximizes their gain but optimizes the system's performance, too. The system is optimal and becomes balanced if it is constructed as a Nash equilibrium, so the problem is to find sequences of utility-improving moves that lead to a Nash equilibrium, starting from some given assignment of resources to agents. The contribution of this paper represents a parallel comparison of $2$ types of methods in resource allocation games that we consider to be suitable to our problem: non-cooperative versus cooperative strategies.