Abstract :
The capacity regions of two new classes of deterministic relay channels are established. In the first class of deterministic relay channels, the family of conditional probability distributions describing the relay channel can be written as p(y2, y3|x1, x2)=p(y3|x1, x2)p(y2|s, x2, y3) where s is a deterministic function of x1, i.e., s=f1(x1). In addition, we require that Srarr(X2, Y2)rarrY3 form a Markov chain for all input probability distributions p(x1, x2). In the second class of deterministic relay channels, there is causal noiseless feedback from relay to sender and the relay output is a deterministic function of x1, x2, and y3, i.e., y2=f3(x2, x2, y3). We consider two alternative schemes to achieve the capacity. The first is based on a generalized strategy of Gabbai and Bross. The second strategy is based on a ldquohash-and-forwardrdquo scheme by Cover and Kim.
Keywords :
Markov processes; channel capacity; Bross strategy; Gabbai strategy; Markov chain; conditional probability distributions; deterministic relay channel capacity; hash-and-forward scheme; probability distributions; Channel capacity; Codes; Decoding; Degradation; Drives; Output feedback; Probability distribution; Relays; Time sharing computer systems; Upper bound;
