| In recent years, the particle swarm optimization (PSO) algorithm and geneticalgorithm (GA) have been extensively used in different research fields and play a veryimportant part as optimizing tools. But, at the same time, they are prone to get trappedin the local optima or have slow convergence rate when complex multimodal problemsare optimized. A new improved hybrid algorithm and a new improved PSO algorithmare designed in order to enhance the reliability and effectiveness of them. Both of themare used to optimize electromagnetic device design by combining Matlab program andHFSS solver. The main works in this paper are listed as follows:1. A hybrid algorithm is proposed which combines basic GA and PSO algorithmwith different parameters. In order to illustrate the effectiveness of the algorithm inrepresentative topology problems, several benchmark functions are used to test theperformance of the improved hybrid algorithm. The results show that it is better than theconventional GA algorithm for multimodal and unimodal functions.2. Combined with the HFSS solver, an ultra-wideband (UWB) antenna and aninterdigital microstrip filter are designed by using the proposed hybrid algorithm,respectively. Through optimizing process, we get the global optimized models and bothof them satisfy the design target.3. In order to enhance the reliability and effectiveness of PSO，an improved PSOalgorithm, neighborhood-redispatch (NR) PSO algorithm, is presented to conquer someshortcomings，such as local optima and local research disability. Several benchmarkfunctions are tested to illustrate the effectiveness of the new algorithm, and the resultsshow that it is better than the conventional PSO algorithm.4. Combined with the HFSS solver, two UWB antennas are designed andoptimized by the proposed NR-PSO algorithm. They not only cover the UWB band(3.1-10.6GHz) and Bluetooth (2.385-2.490GHz) pass-band, but also realize astop-band (5.15-5.90GHz) to avoid potential interferences. |
PDF Full Text Request