Distributed Cross-Layer Optimization of Wireless Sensor Networks: A Game Theoretic Approach

This paper proposes a distributed optimization framework for wireless multihop sensor networks base on a game theoretic approach. We show that the cross-layer optimization problem can be decomposed into two subproblems corresponding to two separate layers (the physical and the application layers) of the overall system. By modelling each subproblem as a noncooperative game, we aim to solve the noncovex application-layer rate-allocation and physical-layer power-allocation subproblems in a distributed manner. Further, we prove the existence, uniqueness, and stability of the Nash equilibria for both games under certain sufficient conditions. Finally, we show that by using a set of dual variables as the market prices to coordinate the physical layer supply and the application layer demand, the overall optimization process strikes a right balance between the two layers in an overall cross-layer design.