Causal relay networks with causal side information

In this paper, we consider causal discrete-memoryless relay networks (DMRNs) with causal side information. The networks consist of multiple nodes, each of which can be a source, relay, and/or destination. There are two types of relays in the network, i.e., relays with one sample delay (strictly causal relays) and relays without delay (causal relays) whose transmit signal depends not only on the past received symbols but also on the current received symbol. Also, we assume that some nodes can causally use channel state information (CSI) as side information. For the network, we derive a new cut-set bound, which recovers the classical cut-set bound and our previous cut-set bound for causal DMRNs without side information. Using an example of a fading relay channel, we show that the new cut-set bound can be achieved by a simple amplify-and-forward type relaying.