Markov Decision Process Based Wireless Multicast Opportunistic Routing

There are more and more pervasive applications of Wireless Multicast Network, such as the video conference, the voice transmission and the software updates etc. The traditional wireless multicast routing method is to construct a multicast tree, which may ignore many available links. This paper presents Markov Decision Process Based Multicast Opportunistic Routing Model. Whether each node of the whole network has received a packet stands for different states of a Markov chains. That is to say that Markov state transition is a routing process. This paper provides a new approach for the study of wireless multicast networks routing. In the reward function of the model, the consumption of Request and ACK among the nodes is abundantly considered. In accordance with each state of the network, the network chooses the optimal forwarding nodes which make the reward function of the model maximum. The experiment and the simulation results proves MDP is most effective routing scheme, and meets with the network system´s requirement.