Searching for Multimode Resonator Topologies with Adaptive Differential Evolution.

Area of Science:

• Electrical Engineering
• Electromagnetics
• Computational Physics

Background:

• Microstrip resonators are crucial components in modern communication, radar, and navigation systems.
• Current design processes for resonators and filters are largely manual, relying on expert knowledge and computationally intensive modeling.
• This limits exploration to known topologies, potentially overlooking novel, high-performance designs.

Purpose of the Study:

• To propose and evaluate an automated approach for discovering new microwave resonator topologies.
• To optimize resonator parameters for specific frequency-selective characteristics.
• To explore the largely uncharted space of possible resonator constructions.

Main Methods:

• Utilized a zero-order optimization algorithm, specifically the L-SRTDE algorithm, for parameter tuning.
• Investigated a family of symmetrical resonators constructed from four rectangles.
• Formulated the topology design as an optimization problem, minimizing deviation between desired and modeled frequency-selective characteristics using electrodynamic modeling.

Main Results:

• The automated approach successfully identified novel resonator topologies.
• These newly found topologies were automatically tuned to meet specified frequency-selective properties.
• Some discovered topologies, while different from existing ones, exhibited high-quality performance characteristics.

Conclusions:

• The proposed automated search method effectively discovers and tunes novel microwave resonator topologies.
• This approach expands the design space beyond conventional structures, offering potential for improved device performance.
• The findings suggest that computational optimization can significantly advance the design of microwave devices.