Distributed power and routing optimization in single-sink data gathering wireless sensor networks

This paper addresses a total transmission power minimization problem in single-sink data gathering wireless sensor network. We propose a distributed algorithm for solving the convex problem with partial dual decomposition approach by jointly optimizing the routing and the power allocation. We assume orthogonal multiple access communications under Rayleigh fading. By applying dual decomposition for relaxing the coupling constraint, the optimization problem is decomposed vertically into two independently solvable subproblems: the routing problem in the network layer and the power allocation problem in the physical layer. Furthermore, second-level dual decompositions are performed for distributing the solution process horizontally within each layer. The master dual problem coordinates the whole solution process by introducing the pricing on the link capacities. Gradient projection method is employed to update the primal and dual variables iteratively. Numerical results are provided to show the convergence properties in a static channel and the tracking ability under time-varying Rayleigh channels.