Title: Simulation-based evolutionary method in antenna design optimization
Authors: Li, Yiming
Institute of Communication Studies
Department of Electrical and Computer Engineering
Keywords: Optimization;Genetic algorithm;Numerical electromagnetic method;Antenna;Maxwell's equations;Finite element method
Issue Date: 1-Apr-2010
Abstract: In this paper, a simulation-based optimization method for the design of antenna patterns in mobile broadcasting, multi-bandwidth operation and the 802.11a WLAN is presented. The simulation-based genetic algorithm (GA) is advanced for the antenna design automation with requested specifications. The corresponding cost function in optimization is evaluated by an external numerical electromagnetic (EM) solver, where the communication between the GA and EM solver is implemented with our unified optimization framework (UOF). An A Z-shaped antenna is explored as an example to express the optimization methodology with respect to the specific return loss. Inspired by the scenario of GA for the optical proximity correction in our earlier work, we firstly partition the edges of the antenna into small segments, and then adjust the movements of each segment to construct a newer geometry for the designed antenna with a better return loss. The external EM solver is then performed to calculate the return loss of the newer antenna. The optimized antenna pattern is achieved when the simulated results meet the specific constraints, and then UOF exports the antenna pattern with the better return loss evaluated by the external EM solver. Otherwise, the evolutionary algorithm will enable us to search for a better solution again. UOF presents the capability in the optimization with an external solver. Our preliminary numerical results confirm the robustness and efficiency of the developed simulation-based optimization method. (C) 2009 Elsevier Ltd. All rights reserved.
URI: http://dx.doi.org/10.1016/j.mcm.2009.08.017
ISSN: 0895-7177
DOI: 10.1016/j.mcm.2009.08.017
Volume: 51
Issue: 7-8
Begin Page: 944
End Page: 955
Appears in Collections:Articles

Files in This Item:

  1. 000274587000013.pdf