An energy efficient topology control scheme with connectivity learning in wireless networks

In wireless networks, due to the variation in environmental and link characteristics, the network topology will change over time. The foremost feature that affects the connectivity and lifetime of a network is the distributed topology control. Nodes in a wireless network are resource constrained. Topology control algorithms should be helpful to improve the energy utilization, reduce interference between nodes and extend lifetime of the networks operating on battery power. This paper proposes a topology control and maintenance scheme while learning the network link characteristics. The system learns the varying network link characteristics using reinforcement learning technique and gives an optimal choice of paths to be followed for packet forwarding. The algorithm calculates the number of neighbors a node can have, which helps to reduce power consumption and interference effects. The algorithm also ensures strong connectivity in the network so that reachability between any two nodes in the network is guaranteed. Analysis and simulation results illustrate the correctness and effectiveness of our proposed algorithm.