We consider a basic cooperative network that is composed of three nodes which use Amplify-and-Forward as a cooperative strategy. This simple setup assumes that each time, one node is the source, one node is the relay, and one node is the destination. Here, we focus on the scheduling problem, which is defined as the distribution of the three roles (source, relay, and destination) to the three nodes. Based on two extreme scheduling approaches, which are a periodical and an instantaneous channel-based distribution of the roles, we propose a number of scheduling algorithms. These algorithms are classified according to three essential design features: fairness, channel adaptation, and distributed coordination. These features are related to the equilibrium, the performance, and the computational complexity of a scheduling algorithm. Among the approaches presented, there is one which jointly satisfies all of them and, thus, is the most appropriate. This solution combines a "smart" round-robin source selection with a normalized instantaneous signal-to-noise ratio relay selection and seems to be a practical solution for ad hoc systems. The proposed algorithms are compared, and their advantages/disadvantages are obtained by analytical and simulation results.