Στοχαστικές πολυδιάστατες τεχνικές βελτιστοποίησης για τη σχεδίαση δικτύων ασύρματων επικοινωνιών

Abstract

Over the last years, the application of stochastic optimization techniques in order to facilitate the solution of problems that involve network planning of wireless systems has become very popular. The main concept behind this approach is their ability to offer quickly near to optimal solutions for complex multi-objective and constrained problems. In the framework of this thesis, the process of 3G network planning and optimization is initially examined, along with the utilization of advanced techniques, such as smart antennas and MIMO systems. A realistic simulator has been developed for this reason, which comprises the main core of the planning and radio resource management process. The presented results indicate that network performance can improve significantly, while the trade-off balance between multiple and conflicting objectives is adjusted in a controllable manner. Therefore, two basic problems of terrestrial digital video broadcasting (DVB-T) network planning are analyzed and solved. Specifically, the procedure of allocating frequency channels to the defined service areas is firstly conducted with the implementation of stochastic algorithms. The transition from analogue to digital television is then investigated in detail. In any case, when the proper mathematical model is provided, the stochastic optimization algorithms prove to be a very powerful tool for the computationally complex and challenging process of network planning.