Equilibrium dynamic systems intelligence

  • Authors:
  • Marius-Constantin Popescu;Onisifor Olaru;Nikos Mastorakis

  • Affiliations:
  • Faculty of Electromechanical and Environmental Engineering, University of Craiova, Craiova, Romania;Faculty of Engineering, University of Constantin Brancusi, Craiova, Romania;Technical University of Sofia, Bulgaria

  • Venue:
  • WSEAS Transactions on Information Science and Applications
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Most work in Artificial Intelligence reviews the balance of classic game theory to predict agent behavior in different positions. In this paper we introduce steady competitive analysis. This approach bridges the gap between the standards of desired paths of artificial intelligence, where a strategy must be selected in order to ensure an end result and a balanced analysis. We show that a strategy without risk level is able to guarantee the value obtained in the Nash equilibrium, by more scientific methods of classical computers. Then we will discuss the concept of competitive strategy and illustrate how it is used in a decentralized load balanced position, typical for network problems. In particular, we will show that when there are many agents, it is possible to guarantee an expected final result, which is a 8/9 factor of the final result obtained in the Nash equilibrium. Finally, we will discuss about extending the above concept in Bayesian game and illustrate its use in a basic structure of an auction.