Abstract
A modifled particle swarm optimization (PSO) algorithm applied to planar array synthesis considering complex weights and directive element patterns is presented in this paper. The modern heuristic classical PSO scheme with asynchronous updates of the swarm and a global topology has been modifled by introducing tournament selection, one of the most efiective selection strategies performing in genetic algorithms the equivalent role to natural selection, and elitism. The modifled PSO proposed combines the abilities of the classical PSO to explore the search space and the pressure exerted by the selection operator to speed up convergence. Regarding the optimization problem, the synthesis of the feeds for rectangular planar arrays consisting of microstrip patches or subarrays of microstrip patches is considered. Results comparing the performance and limitations of classical and modifled PSO-based schemes are included considering both test functions and planar array complex synthesis to best meet certain far-fleld radiation pattern restrictions given in terms of 3D-masks. Finally, representative synthesis results for sector antennas for worldwide interoperability for microwave access (WiMAX) applications are also included and discussed.
Highlights
The particle swarm optimization (PSO) algorithm, based on the movement and intelligence of swarms, has become an attractive alternative to other heuristicLanza, Perez, and Basterrechea approaches such as genetic algorithms (GA), simulated annealing (SA) or ant colony optimization (ACO), and has been successfully applied in different research areas [1,2,3,4,5,6,7,8,9]
The well-known Griewank, Rastrigin, Rosenbrock and Sphere test functions summarized in Table 1, with a zero-value global minimum [24], have been used as the test bed to compare the performance of both classical PSO with asynchronous updates of the swarm and a global topology (PSO hereinafter) and the same scheme modified by introducing tournament selection (TS) and elitism
The results obtained have been appropriately averaged to compare PSO and mPSO schemes, considering parameters such as the success rate (SR), representing the percentage of runs that converge to a valid solution, i.e., those runs for which the value of the function f (x) reaches the specified value to reach (VTR), f (x) < VTR, with a maximum of 300000 fitness function calls allowed; and the average number of fitness function evaluations required to reach the VTR, N Favg, computed considering only the successful runs
Summary
The PSO algorithm, based on the movement and intelligence of swarms, has become an attractive alternative to other heuristic. This work focuses on the analysis of a modified PSO-based approach proposed by the authors to improve the performance of classical schemes. The PSO approach proposed benefits from some characteristics of GA, introducing a selection operator to direct and speed up the search, the tournament selection strategy, along with elitism, applied to ensure that the best particle within the swarm is preserved iteratively [2]. Classical and hybrid PSO schemes, [18] and [19] respectively, have been applied by the authors to linear array synthesis. Progress In Electromagnetics Research, PIER 93, 2009 a wide range of results, comparing in Section 4.1 the performance of both classical and modified PSO schemes using either well-known test functions or a canonical planar array synthesis problem; and summarizing in Section 4.2 representative far-field radiation pattern synthesis results for sector antennas for WiMAX applications.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
