Cognitive radio networks that allow dynamic spectrum access are considered spectrally more efficient than networks using fixed spectral allocation. These networks are characterized by dynamically changing channel sets at each node. The graph theoretic approach used in traditional multi-hop networks fails to efficiently model multi-hop cognitive radio networks and capture the required information for optimal routing. Hence, conventional graph-based routing protocols such as DSR or AODV cannot be used directly, for route discovery in such networks. In this paper, a unique multi-edge planar graph model for routing in such type of networks is proposed, which is quite simple and could be used in conjunction with any conventional graph-based routing protocol. The model is validated through simulations and the complexity of the model is shown to be lesser than an earlier layered graph model. Categories and Subject Descriptors C.2.1 [Computer-Communication Networks]: Network Architecture and ...