7. Conclusions
We develop a new and effective channel assignment method that improves the frequency channel utilization of MRMC WMNs at the expense of increased hardware complexity by incorporating beamforming directly into the conflict graph and matrix during interference modeling. LAB-CA significantly reduces the number of frequency channels required to ensure interference-free communication among the mesh nodes for achieving maximum network throughput. The experimental results show that LAB-CA signifi- cantly outperforms classical OAP-CA in terms of NCR. We extend our channel assignment framework to incorporate heterogeneous mesh nodes in order to model a more realistic WMN architecture. The extended channel assignment method LABCA_HT significantly outperforms OAP-CA_HT in terms of NCR in sparse as well as dense mesh networks. Compared to sparse mesh networks, the throughput per node is lower and the solution times of the routing stage are higher in dense mesh networks. The throttling of the throughput per node in dense mesh networks can be alleviated by adding more gateways. Also, when using multiple gateways, mesh nodes in a dense mesh network can be divided nto groups, one group per gateway. This will divide the computational effort among the multiple gateways, which will reduce the solution times. As part of our future work, we plan to incorporate multiple gateways in our channel assignment framework.
