Entropy-Based Estimation of Client Preferences in Wireless Push Systems

Knowledge of the clients' preferences in communication systems enables adaptive, precise resource allocation in real-time. The present work introduces a novel scheme for estimating the clients' preferences in wireless data broadcast systems. The proposed scheme relies on the principle of maximum entropy, with statistical moments of the clients' overall waiting time serving as inputs, and the probability distribution of the broadcast data items being the output. Observations regarding periodic broadcast scheduling enabled the reduction of the convergence time well below that of related well-known approaches, in both noisy and noiseless environments. Simulation results showed that the technique can be used for accurate adaptation in minimal times, compared to the alternatives.