Title :
Application of evolutionary computation techniques to nonlinear microwave circuit analysis
Author :
Iezekiel, Stavros
Author_Institution :
Sch. of Electron. & Electr. Eng., Leeds Univ.
Abstract :
The application of genetic algorithms to the simulation of nonlinear microwave one-port circuits using hybrid techniques is reviewed. To date, harmonic balance and sample balance have been the two major types of hybrid technique, and genetic algorithms have been successfully used to implement both methods. However, genetic algorithms are “transparent” to the solution domain, which allows new classes of hybrid technique to be developed: these are termed frequency domain sample balance (FDSB) and time-domain harmonic balance (TDHB). FDSB, in particular, has useful properties for circuits with noncommensurate multitone inputs. By formulating the error function in the time domain whilst representing the solution in the frequency domain, it is possible to circumvent the use of Fourier transforms altogether. This allows, in principle, any arbitrary steady-state input to be applied
Keywords :
circuit optimisation; circuit simulation; frequency-domain analysis; genetic algorithms; microwave circuits; nonlinear network analysis; time-domain analysis; error function; evolutionary computation; frequency domain sample balance; genetic algorithm; hybrid analysis technique; noncommensurate multitone excitation; nonlinear microwave one-port circuit analysis; steady-state simulation; time domain harmonic balance; Circuit simulation; Computational modeling; Evolutionary computation; Fourier transforms; Frequency domain analysis; Genetic algorithms; Microwave circuits; Microwave theory and techniques; Steady-state; Time domain analysis;
Conference_Titel :
High Performance Electron Devices for Microwave and Optoelectronic Applications, 2000 8th IEEE International Symposium on
Conference_Location :
Glasgow
Print_ISBN :
0-7803-6550-X
DOI :
10.1109/EDMO.2000.919064