Using learning automata for adaptive push-based data broadcasting in asymmetric wireless environments

Abstract

Push systems are not suitable for applications with a priori unknown, dynamic client demands. This paper proposes an adaptive push-based system. It suggests the use of a learning automaton at the broadcast server to provide adaptivity to an existing push system while maintaining its computational complexity. Using simple feedback from the clients, the automaton continuously adapts to the client population demands so as to reflect the overall popularity of each data item. Simulation results are presented that reveal the superior performance of the proposed approach in environments with a priori unknown, dynamic client demands.

Query Management for Efficient Data Access in Wireless Environment

“…Although, such a priori knowledge seems like a strong assumption, there are several methods for determining the data access probabilities. [10,11,12,13]. One of the methods to count access probability is to gather data access statistics at regular interval of time and draw inference from it.…”

On regional performance improvement of an adaptive wireless push system in environments with locality of demand

“…In the adaptive wireless push system [6], which enhanced the non-adaptive one of [4], the server is equipped with an Smodel Learning Automaton that contains the server's estimate p i of the demand probability d i for each data item i among the set of the items the server broadcasts. Clearly,…”

“…It is clearly seen, that convergence of the item probability estimated by the automaton to the overall client demand for this item is achieved. Moreover, simulation results in [6] and [18] have demonstrated efficient operation in environments characterized by dynamic and a-priori unknown to the server, client demands.…”

“…The proposed approach can trade the E b /N 0 redundancy at clients in groups at distances d < R for an increased bit rate for the broadcast of the items demanded by these groups. Knowledge of client positions is conveyed to the server via a simple feedback pulse from the clients, a mechanism that was used in [6] in order to provide adaptivity to dynamic client demands. Additionally, a simple mechanism is introduced that protects performance around the geographical area of interest from degradation caused by clients that are located elsewhere and demand the same information items with clients inside that area.…”

“…Simulation results which reveal the performance superiority of the proposed approach to that of the fixed rate adaptive wireless push scheme of [6] in environments with locality of demand are presented in Section III. Finally, Section IV summarizes and concludes the paper.…”