A Scalable Multiagent Approach for Channel Assignment in Wireless Networks T.L.Monteiro DAELN, Federal Technological University of Parana Curitiba, Brazil tmonteiro@utfpr.edu.br M.E.Pellenz PPGIa, Pontifical Catholic University of Parana Curitiba, Brazil marcelo@ppgia.pucpr.br E.Jamhour PPGIa, Pontifical Catholic University of Parana Curitiba, Brazil ejamhour@gmail.com M.C.Penna PPGIa, Pontifical Catholic University of Parana Curitiba, Brazil penna@ppgia.pucpr.br F.Enembreck PPGIa, Pontifical Catholic University of Parana Curitiba, Brazil fabricio@ppgia.pucpr.br R.D.Souza CPGEI, Federal Technological University of Parana Curitiba, Brazil richard@utfpr.edu.br ABSTRACT The performance of IEEE 802.11-based wireless local area net- works (WLANs) depends on the channel assignment among in- terfering access points (APs). Due to the limited number of non- overlapping channels, severe interference scenarios may arise if no appropriated spectrum management is employed. Additionally, in dense urban areas it is usual to find wireless networks scenarios with interfering APs belonging to different administrative domains. In such cases the use of centralized algorithms is not feasible and the already proposed distributed methods present scalability prob- lems. Our study investigates the channel allocation problem in WLANs, aiming to minimize interference and improve the network performance. In this paper, we formalize the channel allocation as a distributed constraint optimization problem (DCOP) and propose a new cooperative channel allocation strategy, called Distributed Suboptimal Channel Assignment (DSCA), which uses distributed pseudotree-optimization (DPOP). Our strategy reduces the number of messages and also the total amount of control information ex- changed between APs. The results show that DSCA outperforms known techniques in terms of scalability, while keeping a good so- lution quality. Keywords Wireless Networks; Channel Assignment; Distributed Constraint Optimization; ADOPT. 1. INTRODUCTION The efficient deployment of wireless local area networks (WLANs) becomes a challenge, due to the widespread use of IEEE 802.11- based networks. Although these networks promise high-speed ac- Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. SAC’16, April 4-8, 2016, Pisa, Italy Copyright 2016 ACM 978-1-4503-3739-7/16/04. . . $15.00 http://dx.doi.org/xx.xxxx/xxxxxxx.xxxxxxx cess, many factors contribute to lower the effective rates [1]. An important source of performance degradation is the interference caused by neighboring access points (APs) operating on overlap- ping channels. The level of interference depends mainly on chan- nel allocation, the geographical distribution of APs and clients, the number of clients associated with each AP and the traffic load. The IEEE 802.11 standards operating in the 2.4GHz ISM band pro- vide a set of 11 frequency channels, but only three channels can be used simultaneously by neighboring APs without interference (non-overlapping channels) [2, 3]. The ability to use these channels to minimize interference can substantially increase the effective use of available bandwidth and network capacity [4]. In many practical scenarios, like residential WLAN deployment in dense urban areas, the interfering APs belong to different admin- istrative domains, thus a cooperative channel assignment strategy is necessary. Some cooperative strategies require the exchange of control information through a wired network [5], what is not fea- sible when considering APs from different administrative domains. It is also possible to apply uncoordinated strategies [6], but it may require explicit client cooperation, what imposes additional pro- cessing and traffic overhead. Several channel allocation algorithms have been proposed in the literature, based on a wide range of different techniques, including game theory [8], auction mechanisms [9], local bargaining [10] and graph coloring [12]. However, these algorithms are not suitable for distributed implementation, what is a requirement when APs be- long to different administrative domains. Therefore, in the follow- ing, we discuss the proposals that use distributed and cooperative approaches for channel allocation in WLANs. In [13], each AP monitors packet transmissions, searching for the most lightly loaded channel, and switches to operate on it until the next scan finds a less congested one. The proposed strategy uses traffic-related information to select a channel for an AP. However, the algorithm fails to capture situations when clients associated to two APs interfere with each other, while the APs do not interfere among themselves. For this reason, the algorithm cannot take ad- vantage of channel re-use opportunities. In [2], the authors formulate channel assignment as a weighted vertex-coloring problem and presents two distributed algorithms.