Joint-optimal probing and scheduling in wireless systems

Consider a wireless system where a sender can transmit data to various users with independent and varying channel conditions. To maximize its long-term transmission rate, the sender should always transmit to the user with the best channel. To discover which user has the best channel, it has to spend time to probe channels, and this reduces the time available for effective transmission. This paper aims at identifying optimal joint probing and scheduling strategies. These strategies realize the best trade-off between the channel state acquisition and effective transmission. We first provide general structural properties of optimal strategies, and then exactly characterize these strategies in particular but relevant cases. Finally we propose extensions of this problem, e.g., to impose fairness among the users, we investigate how to maximize system utility rather than throughput.