Broadband HF Antenna Matching Network Design Using a Real-Coded Genetic Algorithm

摘要

A real-coded genetic algorithm was successfully applied to the matching network problem. Our approach reduces the search across different network topologies by the means of a generalized network description, in which the algorithm "decides" whether a reactive element is inductive or capacitive. Compared to previous analytical and semianalytical techniques that do not require a priori information about the network topology, this method is considerably simpler and its application is straightforward, regardless of the network complexity. Our approach accounts for constrained lossy elements, tolerances, and network on-site tuning. The results for a canonical load show the effectiveness of this method in comparison with other known methods. Finally, for HF broadband antennas, which involve network complexity limits and component value restrictions, it will be shown that the use of lossy networks may help to fulfil system requirements, with only a moderate impact on system efficiency