Multiobjective Optimal Antenna Design Based on Volumetric Material Optimization

Abstract

There is growing interest for small antennas that concurrently have higher functionality and operability. Multiobjective optimization is an important tool in the design of such antennas since conflicting goals such as higher gain, increased bandwidth, and size reduction must be addressed simultaneously. In this paper, we present a novel optimization algorithm which permits full volumetric antenna design by combining genetic algorithms with a fast hybrid finite element-boundary integral method. To our knowledge, this is the first time that a full three dimensional antenna design is pursued using concurrent shape, size, metallization as well as dielectric and magnetic material volume optimization. In comparison to previous optimization pursuits, our approach employs a wide-frequency sweep using a single geometry model, thus, enhancing speed, along with several discrete material choices for realizable optimized designs. The developed algorithm can be interpreted as a three dimensional Pareto optimization scheme and provides the designer with several choices among the best antennas, according to design goals and constraints. The final designs are associated with very thin ( 0 01 ) material substrates and yield as much as 15% bandwidth using a 0 1 -0 4 aperture subject to various gain and bandwidth requirements.Index Terms-Pareto surface font, printed conformal antennas, size shape and material optimization.

Topology optimization of metallic devices for microwave applications

“…In [14] an optimization scheme based on genetic algorithms (GA) showed a way to circumvent the skin depth issue in FEM. The method is restricted to patch antennas and uses a perfect electric conductor (PEC) condition between and on top of the dielectric layers of the patch antenna, to model the conductor.…”

Design of a metamaterial-based linear insulated antenna using a genetic algorithm

“…They have been shown to be useful for optimizing sidelobe levels in phased arrays [16,17] and for optimizing element spacings in Yagi-Uda antennas [18], among other applications [19]. There are certainly other optimization methods that have been recently discussed in the literature that could be employed for this problem, such as multiobjective optimization [20,21] and particle swarm optimization [22], but our goal here is not so much to highlight a method of optimization as it is to highlight the resulting design of the antenna. So, we prefer the more conventional GA for its simplicity.…”

Recent developments in simulation-driven multi-objective design of antennas

“…However, real-world antenna design tasks are multi-objective ones. In particular, if the designer priorities are not clearly defined beforehand, identifying a set of alternative design representing the best possible trade-offs between conflicting objectives may be of fundamental importance (e.g., in order to determine limitations of a given antenna structure and its suitability for a given application) [16][17][18][19]. Nowadays, population-based metaheuristics are undoubtedly the most popular solution approaches for handling multi-objective antenna design problems.…”

“…Nowadays, population-based metaheuristics are undoubtedly the most popular solution approaches for handling multi-objective antenna design problems. Techniques such as multi-objective genetic algorithms (GAs) and particle swarm optimizers (PSO), e.g., [16,[18][19][20][21][22][23], allow finding the entire Pareto front in one algorithm run. However, their disadvantage is high computational cost (hundreds, thousands or even tens of thousands of objective function evaluations), which becomes a serious bottleneck if high-fidelity discrete EM simulations are involved in antenna evaluation process.…”