توزیع یادگیری استراتژیک و فرمول بندی نظری بازی از کدگذاری جاسازی شده شبکه

Nowadays, there is a growing demand of middleware solutions for reliable event notification over the Internet due to several key industrial projects aiming at integrating existing legacy systems and enhancing their functionalities. Reliable event notification is mainly realized by means of retransmissions, with the consequent worsening of the performance and traffic load. Traditionally, the opposite approach of spatial redundancy is not used due to the tunability, scalability and flexibility issues of its implementation schemes. In this paper, we propose a game theoretic formulation of Forward Error Correction (one of the mean schemes to introduce spatial redundancy), in order to resolve these mentioned issues. Moreover, we introduce distributed strategic learning for the optimal formulation of the payoff functions in the game, and for the effective adaptivity in response to the possible variations in the experienced loss patterns. We prove the quality of this solution by using a series of simulations run on OMNET++.