A variable neighborhood search method for multi-objective channel assignment problem in Multi-Radio WMNs

Channel assignment schemes in Multi-Radio Wireless Mesh Networks (MR-WMNs) usually leave several links sharing the same channel within overlapped transmissions or interference ranges; this is especially true when only one radio is used or when the number of radios is very small compared to the number of orthogonal channels. In this paper, we propose a new multi-objective optimization model for channel assignment (CA) performed during the MR-WMNs planning process. Given the expected traffic demand, the goal is to (1) minimize user handoff overhead; (2) minimize traffic load variances to achieve load balancing; (3) maximize overall throughput; and (4) maximize Jain´s fairness index to achieve fairness among mesh clients. We also propose a variable neighborhood search (VNS) meta-heuristic to solve our model. Simulation results show that our proposed approach achieves good performance in terms of delay, loss rate, overall throughput and fairness in the MR-WMNs.