Date Added: Apr 2012
In this paper, an optimal algorithm of spectrum decision making is presented for a real cognitive radio network with tree-based topology. All nodes of a subnet in such network have the capability in being aware of spectrum information by both energy detector based sensing and centralized cooperative sensing. After gathering sensed information, the master node will decide which frequency can be used by the subnet and which slave nodes should leave the subnet if there is no common frequency among all nodes. The problem is how to keep nodes staying in the subnet as many as possible.