Use of Genetic Algorithm and Evolution Strategy when Revealing the Worst Case Effects of Crosstalk Propagation in PCB Bus of Spacecraft Autonomous Navigation System

摘要

Importance of the genetic algorithm (GA) and evolution strategy (ES) usage in the investigation of an ultrashort pulse peak voltage in a printed circuit board (PCB) bus of autonomous navigation system (ANS) is highlighted. Trapezoidal ultrashort pulse propagation along the conductors of the PCB bus was optimized. The optimization was made by maximization criteria of maximum crosstalk amplitude at the preset point. The two methods were used for the optimization, which results are compared. The ES optimization was run 20 times when the initial solution was 300 ps. The GA optimization was run 10 times with the parameters: the number of chromosomes ?5, 10; the number of populations ?5, 7, 10, 15; mutation coefficient ?0.1; crossover coefficient ?0.5. After the ES and GA optimizations the crosstalk maximums of 320% and 78% of steady state level in the active conductors respectively were revealed and localized.