Deployment of Large-Scale WLANs Jovan Stosic 1 , Zoran Hadzi-Velkov 2 and Liljana Gavrilovska 2 1 Makedonski Telekomunikacii, Orce Nikolov bb, 1000 Skopje, jovan.stosic@mt.com.mk 2 Ss. Cyril and Methodius University, Faculty of Electrical Engineering, Karpoš 2 bb, 1000 Skopje, {zoranhv | liljana}@etf.ukim.edu.mk Abstract– This paper analyzes the deployment issues for large-scale WLANs and presents our latest results on indoor propagation modeling and network planning. Proper network planning is necessary for large WLAN installations in order to achieve adequate coverage, and it relies heavily on the propagation model. We used the dominant path method to predict the propagation loss for each possible reception point in an indoor environment. Based on this propagation model, we further examined different combinatorial optimization methods to obtain close to optimal positioning of the WLAN access points and compare their cost effectives to the simple installation methods. The optimization algorithms evaluate an objective function that aims to maximize both the coverage area and the overall signal quality over a discrete search space. We propose a combination of two algorithms, Genetic Algorithms or Simulated Annealing, for the initial set of base stations positions, followed by Pattern Search algorithm, for the final accurate positions. Keywords – WLAN, radio propagation modeling, network planning and optimization, simulated annealing, genetic algorithms 1. INTRODUCTION Wireless LANs are already widespread in home and office environments, providing best-effort services at high-data rates while supporting limited user mobility. The increase of the density of the Access Points (AP) in a given indoor environment stimulates the need of their proper deployment to achieve adequate coverage. The traditional approach to cellular network deployment relies on advanced coverage and capacity planning to achieve optimal infrastructure with minimal number of base stations while sustaining a given quality of service. The WLANs, however, are designed to provide low-cost connectivity and are usually deployed in an ad-hoc fashion. Advanced coverage planning is regarded as too complex and too costly for WLANs, hence we investigate ways to reduce complexity and improve WLAN performance by different types of network planning [1]. While deploying WLANs for a small home network might be easy, deploying them for a large enterprise is a non-trivial task. The overall complexity of the network planning problem further depends on the number of AP sites to be optimized and the frequency at which the network operates. By increasing the frequency, the achievable cell size shrinks; hence, more APs are required to cover the same area. The simplest way to install a WLAN is to skip propagation analysis and install a few APs in the areas where coverage is required and easy access to a wired backbone infrastructure is available. It does not require a specific radio network planning, so this method is most likely to be employed by the end users themselves. Although simple, this deployment method may result in a wireless network with coverage gaps in some areas. A second approach is to divide the service area (i.e., the part of a building in which wireless access should be provided) into K equally sized rectangles, where K is the number of available APs, and install APs in the center of each rectangle. In such a grid installation, the AP sites are uniformly distributed within the service area, thus reducing the probability of coverage gaps. The coverage optimization is the most complex method, as it involves the use of propagation analysis and optimization algorithms for determining the optimal AP positions that provide adequate coverage with a minimum infrastructure density. Coverage optimization can reduce the number of required APs, but the potential cost saving can be limited due to the relatively low costs of WLAN equipment, thus necessitating a careful cost trade-of. The remainder of this article is organized as follows. We first discuss the specific problems involved in the deployment of WLANs and preset three common deployment approaches. In Section 2 we give a review of commonly used propagation models and present the used ones. Comparison of the user deployment model to coverage optimization method is given in third section. Section 4 presents the used objective function, review a number of optimization algorithms for objective minimization, and selects the most suitable algorithm. We then conclude the article.