Myopic coding in wireless networks

We investigate the achievable rate of data transmission from sources to sinks through a multiple-relay network. We study achievable rates for omniscient coding, in which all nodes are considered in the coding design at each node. We find that, when maximizing the achievable rate, not all nodes need to “cooperate” with all other nodes in terms of coding and decoding. This leads us to suggest a constrained network, whereby each node only considers a few neighboring nodes during encoding and decoding. We term this myopic coding and calculate achievable rates for myopic coding. We show by examples that, when nodes transmit at low SNR, these rates are close to that achievable by omniscient coding, when the network is unconstrained . This suggests that a myopic view of the network might be as good as a global view. In addition, myopic coding has the practical advantage of being more robust to topology changes. It also mitigates the high computational complexity and large buffer/memory requirements of omniscient coding schemes.