Affordable Access

Publisher Website

Mono- and multiobjective formulations for the indoor wireless LAN planning problem

Computers & Operations Research
Publication Date
DOI: 10.1016/j.cor.2007.02.011
  • Wireless Networks Planning
  • Tabu Optimization
  • Multiobjective Optimization
  • 802.11
  • Computer Science
  • Engineering


Abstract Wireless LANs (WLANS) experienced great success in the past five years. This technology has been quickly adopted in private and public areas to provide a convenient networking access. The fast pace of development has often induced an uncoordinated deployment strategy where WLAN planning tools have been barely used. This article highlights the difficulty of planning such wireless networks for indoor environments. The first issue the WLAN planning problem has to face is to accurately describe the quality of a network, based on realistic propagation predictions. The second issue is to implement a search strategy that provides efficient deployment strategies. This article is introduced by a description of previously proposed planning strategies. Their study opens out onto a problem formulation that accounts for coverage, interference level and quality of service (in terms of data throughput per user). This formulation is then introduced as either a mono- or a multiobjective (MO) optimization problem. In the first case, we propose to solve the mono-objective problem with a Tabu search metaheuristic minimizing a weighted sum of the planning criteria. Then, we compare the outcome of this strategy to the results of our previously proposed MO Tabu search strategy. We highlight the fact that efficient solutions are obtained quickly with the mono-objective approach if an appropriate set of weighting coefficients of the evaluation function is chosen. The main issue of mono-objective search is to determine these coefficients. It is a delicate task that often needs several runs of the algorithm. MO search is an interesting alternative heuristic as it directly provides a set of planning solutions that represent several trade-offs between the objectives. Our MO heuristic looks for a set of non-dominated solutions expected to converge to the Pareto front of the problem and selects the most significant ones for the end user. Both QoS-oriented planning methods are illustrated on a realistic environment representing a building floor of about 12 600 m 2 . Results show the assets of both approaches but mainly emphasize the benefit of the MO search strategy that offers several alternative solutions to the radio engineer.

There are no comments yet on this publication. Be the first to share your thoughts.