Partner selection for decode-and-forward cooperative relaying: A matching theoretic approach

Abstract : A matching theoretic approach to study the partner selection in cooperative relaying is followed. Partner selection is considered as a special stable roommate problem where each player ranks its partners by some criterion. Each agent aims here at finding a "good" partner in order to exploit efficiently the spatial diversity achieved with cooperation. We adapt Irving's algorithm [3] for determining the partners of each player. The ranking criterion here is chosen to be outage probability such that each player comprises its own preference list according to outage probability from the lowest to the highest. The first player in the preference list provides the lowest outage probability. We introduce a decentralized version of Irving's algorithm. Then, we compare the results obtained by stable-matching with the global optimum and random selection results. From the computational results, we observe that stable-matching results are near to global optimum as well as superior than random selection in terms of average outage probability.