Efficient citywide planning of open WiFi access networks using novel grouping harmony searchheuristics

  • Authors:
  • I. Landa-Torres;S. Gil-Lopez;J. Del Ser;S. Salcedo-Sanz;D. Manjarres;J. A. Portilla-Figueras

  • Affiliations:
  • TECNALIA, Telecom unit, Parque Tecnológico de Bizkaia, Ibaizabal Bidea, Edificio 202, Zamudio, Bizkaia, Spain;TECNALIA, Telecom unit, Parque Tecnológico de Bizkaia, Ibaizabal Bidea, Edificio 202, Zamudio, Bizkaia, Spain;TECNALIA, Telecom unit, Parque Tecnológico de Bizkaia, Ibaizabal Bidea, Edificio 202, Zamudio, Bizkaia, Spain;Department of Signal Theory and Communications, Universidad de Alcalá, Spain;TECNALIA, Telecom unit, Parque Tecnológico de Bizkaia, Ibaizabal Bidea, Edificio 202, Zamudio, Bizkaia, Spain;Department of Signal Theory and Communications, Universidad de Alcalá, Spain

  • Venue:
  • Engineering Applications of Artificial Intelligence
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper proposes the application of a novel meta-heuristic algorithm to the metropolitan wireless local area network deployment problem. In this problem, the coverage level of the deployed network must be maximized while meeting an assigned maximum budget, set beforehand. Specifically, we propose an approach based on the Harmony Search (HS) algorithm, with three main technical contributions: (1)the adaptation of the HS algorithm to a grouping scheme; (2)the adaptation of the improvisation operators driving the algorithm to the specific characteristics of the optimization problem to be tackled; and (3)its performance assessment via a simulated experiment inspired by real statistics in the city of Bilbao (Basque Country, northern Spain). Moreover, a comparison study of the proposed algorithm with a previous published grouping genetic algorithm is carried out, to further validate its performance. In light of the simulation results obtained from extensive experiments and several complexity considerations, we conclude that the proposed algorithm outperforms its genetically inspired counterpart, not only in terms of computation time, but also in the coverage level of the solution obtained.