Particle Swarm Optimization in Comparison with Classical Optimization for GPS Network Design

Abstract

The Global Positioning System (GPS) is increasingly coming into use to establish geodetic networks. In order to meet the established aims of a geodetic network, it has to be optimized, depending on design criteria. Optimization of a GPS network can be carried out by selecting baseline vectors from all of the probable baseline vectors that can be measured in a GPS network. Classically, a GPS network can be optimized using the trial and error method or analytical methods such as linear or nonlinear programming, or in some cases by generalized or iterative generalized inverses. Optimization problems may also be solved by intelligent optimization techniques such as Genetic Algorithms (GAs), Simulated Annealing (SA) and Particle Swarm Optimization (PSO) algorithms. The purpose of the present paper is to show how the PSO can be used to design a GPS network. Then, the efficiency and the applicability of this method are demonstrated with an example of GPS network which has been solved previously using a classical method. Our example shows that the PSO is effective, improving efficiency by 19.2% over the classical method.