This document proposes using a Markov chain model and bipartite graphing to efficiently schedule spectrum in cognitive radio networks. It models the cognitive radio network as a k-connected bipartite graph and uses a Markov chain to represent the state transitions of channels between idle and busy. It then applies the Banker's algorithm to the modeled cognitive radio network to allocate spectrum to users while avoiding deadlock. The proposed approach indicates it could improve spectrum scheduling and allocation performance in cognitive radio networks.