No abstract

No abstract

We propose a new distributed algorithm for optimal channel assignment in WLANs with multiple access points, applying a novel formulation for wireless networks based on Distributed Constraint Optimization Problem (DCOP). The DCOP approach allows to model a wide variety of distributed reasoning tasks of multi-agent applications. The proposed strategy is derived from a polynomial-space algorithm for DCOP named ADOPT, which is guaranteed to find the global optimal solution while allowing agents to execute asynchronously and in parallel. Our proposed algorithm, denoted DCAA-O, allows a group of APs to coordinate themselves in order to find the optimal channel allocation solution which minimizes the network interference. The algorithm performance is evaluated in terms of the required number of transmitted control messages among APs. It is shown that DCAA-O outperforms a recently proposed channel assignment strategy for WLANs.

No abstract

International audienceAn important design goal for wireless local area networks (WLANs) is the performance. It is well known that the major issue challenge to the performance in WLANs is the wireless interference. This leads to an ambience that depends on the channel assignments among interfering access points (AP). Due to the limited number of non-overlapping channels, severe interference scenarios may arise if an unappropriate spectrum planning is considered. In our study we considered WLANs scenarios where APs belong to different administrative domains, a common situation in dense urban deployments. In such environment the use of centralized algorithms is not appropriate and the already proposed distributed methods do no guarantee optimal channel assignment. In this paper, we formalize the channel allocation as a distributed constraint optimization problem (DCOP) and propose a new cooperative channel allocation strategy using the distributed pseudo tree-optimization procedure (DPOP). The adjacent channel interference is analytically formulated for DPOP. The simulation results show that the proposed strategy always achieves the optimal solution and is scalable in terms of the number of exchanged messages