On achievable rates for relay networks based on partial decode-and-forward

In this paper, we investigate the performance of two strategies: simultaneous sliding window decoding (SimSlide) and simultaneous backward decoding (SimBack) when generalizing partial decode-and-forward to the relay networks. We show that these two methods yield the same achievable rates for relay networks but in comparison with sequential decoding, they yield higher rates. This is due to the fact that simultaneously decoding the different parts of the source message yields fewer bounds and this ultimately results a comprehensive and higher rate. We also introduce a special class of semi-deterministic relay networks and obtain its capacity by the proposed achievable rate.