Near-optimal data allocation over multiple broadcast channels

Data broadcast has become a promising solution for information dissemination in the wireless environment. In this system, the average expected delay (aed) depends on the broadcast schedule because different data items may have different popularities and there are multiple broadcast channels for current delivery. In this paper, a restricted dynamic programming approach to minimize the aed is proposed to partition data items over multiple channels near optimally. To reduce costs in dynamic programming, for each partition, we predict a possible location which may be very close to the optimal cut by using a low bound of aed for given items among given channels. Thus, the search space in dynamic programming can be restricted to the interval around the found cut point and only takes O(NlogK) time. Moreover, a valley approximation algorithm is performed to improve the aed. Simulation results show that, the hit rate obtained by our algorithm is higher than 95% and it also outperforms the existing algorithm 200%.