Multi-source-destination distributed wireless networks: Pareto-efficient dynamic power control game with rapid convergence

Abstract

A game-theoretic method for transmit power control across multi-source-destination distributed wireless networks is proposed, which is viable for any number of source-destination pairs, with any number of players (or sources). A dynamic noncooperative repeated game is proposed to optimize both packet delivery ratio (PDR) and transmit power considering a realistic signal-to-interference-plus-noise ratio (SINR) model of the wireless channel. Here, the sources, which are players, transmit concurrently and, thus, have imperfect information about the actions of other players. The game accounts for a limited set of discrete values for transmit power, and the game can be applied in static, quasi-static, and slow-fading channels. If the SINR is feasible, each game stage has a subgame perfect equilibrium, and the game requires fewer iterations to converge to a Pareto-efficient outcome than other appropriate techniques such as SINR discrete power balancing and multiobjective power optimization. In this context, a novel accurate PDR model is given in terms of a compressed exponential function of inverse SINR, which is a function that is realistic for many IEEE 802.11-type implementations of various packet sizes and data rates, and facilitates a tractable analysis and implementation of this dynamic game.