| Popis: |
This thesis addresses itself to the computer-aided design and modelling of microwave circuits using efficient minimax and ℓ₁ optimization techniques. Recent algorithms for nonlinear minimax and ℓ₁ for the particular applications of this thesis. Efficient gradient approximation techniques applicable to both minimax and ℓ₁ optimization are presented. A comparative example with ℓ₂ optimization illustrates the robustness of ℓ₁ for the particular applications of this thesis. Efficient gradient approximation techniques applicable to both minimax and ℓ₁ optimization are presented. A simplified and straightforward treatment of sensitivities for two port networks, cascaded and branched cascaded structures is introduced. The objective is to calculate the sensitivities efficiently, without appealing to the adjoint network concept. A novel proof of a recent result in sensitivity analysis of lossless and reciprocal two-ports is presented. Design of manifold type waveguide multiplexers has been considered as a major application for both minimax optimization and the theoretical work in branched cascaded network sensitivity analysis. Components of the multiplexer structure and nonideal effects such as dissipation and dispersion are discussed and a step-by-step implementation of an algorithm to calculate particular responses and sensitivities is presented. Examples of the design of 3-, 12- and 16-channel, 12 GHz multiplexers illustrate the practicality of the approach presented. A new approach to modelling of microwave devices which exploits the theoretical properties of the ℓ₁ norm is presented. The concept of multi-circuit measurements is introduced and its merits in obtaining unique and self-consistent parameters are discussed. The technique is applied to modelling of multi-coupled cavity filters and GaAs FET's. The application of efficient modelling techniques in developing algorithms for postproduction tuning and in establishing the relationship between physical parameters of a device and its equivalent circuit model parameters is discussed. Doctor of Philosophy (PhD) |
