Indoor distributed antenna systems deployment optimization with particle swarm optimization

Abstract

Ever increasing competition for energy and other resources raises the need for applying optimized autonomous services for the e�cient access to and utilization of these resources. Biologically inspired optimization processes are suitable to achieve balanced distribution of and constrained access to scarce resources. Speci�cally, an optimization mimicking the behavior of a swarm of species displays natural self-adaptation and self-organization. This thesis provides a general framework utilizing bio-inspired techniques to solve one of the major problems in communication networks - optimized deployment cost and minimized power deviation of an In-Building Distributed Antenna System (IB-DAS). The way how these two mutually co-dependent objectives can be solved concurrently with Particle Swarm Optimization (PSO) is reported. IB-DAS extends the wireless access from the base station to distributed antennas through the complex network of coaxial cables and power splitters. For high rise buildings and housing complexes, the initial costs of (IB-DAS), incurred mainly by the network equipment, cabling and splitters, are very high. In turn, the running costs of powering the network could also be high, especially if poor design leads to antennas overpowering. An optimal design of (IB-DAS) that minimizes both of these cost components is therefore critical. Among the results reported on is a novel IB-DAS optimization model that utilizes PSO to provide an optimal network topology, simultaneously minimizing the cost of the cabling and equipment for the whole building as well as minimizing the maximum power deviation among all the antennas. The use of PSO is motivated by the extremely high complexity of the problem. Our model is a complete proposition for (IB-DAS) optimization with demonstrated scalability and one that delivers robust (IB-DAS) designs even for the tallest buildings beyond hundred oors. Numerically conducted case studies for realistic (IB-DAS) layouts show that the proposed model delivers optimal and scalable solutions for small and medium height buildings and near optimal solutions for very tall buildings. The deployment cost increases with a factor of 1.04 -1.06 when power deviation optimization is considered while optimizing the total cost. PSO-DAS model is able to return solutions for tall buildings in maximum 5 minutes unlike GA-DAS model - developed for comparison purposes - that requires in some cases 14 times the time spent by PSO-DAS model to return almost the same solution.

Date of Award	2015
Original language	American English
Supervisor	Abdel Isakovic (Supervisor)