Abstract:

An efficient resource allocation mechanism was proposed for information diffusion in wireless autonomic networks. First, the dynamic arrival and leave process of the information data is characterized by a transmission queue. To cope with the dynamic transmission queues and wireless channels, the information diffusion problem is formulated as a multi-user Markov decision process (MMDP) which can be decomposed by using the principle of dual method. To reduce the computation complexity, a model-based online learning method was proposed based on which the user can make a foresighted decision by iterating only once when the network state changes.

Key words: autonomic network, data diffusion, Markov decision process