Search

Byline: Alexandre Gondran, Alexandre Caminada, Julien Fondrevelle, Oumaya Baala The Wireless Local Area Networks (WLAN) large-scale deployment is a difficult parameterisation work and it requires automatic planning process with formal models. This paper proposes a solution to this matter and provides a model integrating all useful parameters to estimate the location and parameter settings of such network while optimising network costs and throughput. The access point location, the access point pattern, the antenna orientation, the emitted power and the antenna frequency channel are set in this model which brings together cell and frequency planning features (mix of Automatic Cell Planning (ACP) and Automatic Frequency Planning (AFP) problems). With the formal model description this paper illustrates step by step the process of WLAN planning from the AP location to the effective throughput computation. A fair example is given at each section to didactically emphasise the planning problem and the model we defined. At the end, the optimisation problem is to minimise the economical cost of deployment and the economical cost of throughput loss inside a unique and understandable objective function.

Byline: Hakim Mabed, Alexandre Caminada, Jin-Kao Hao Cell capacity planning is a key phase of the expansion of cellular networks. The growth of communications demand requires improving the capacity of the networks regularly. However, cell capacity planning is difficult to achieve while it must take into account two conflicting objectives: minimising the blocking rate and reducing the interference. In TDMA systems, current literature offers few studies dealing simultaneously with both criteria. In this paper, we propose a bicriteria model for TDMA cell capacity planning as answer to a flexible management of the tradeoff between blocking and interference phenomena. While cell capacity dimensioning requires the offered traffic load on each station, we also present a module for offered traffic computation where user retrials and redials are considered. At the end, we carry out tests on two real world datasets with different network architectures and traffic loads in order to assess the efficiency of the approach.