A new adaptive sampling is proposed to accelerate frequency-domain calculations. The algorithm is based on the survival-of-the-fittest principle of genetic algorithms and uses rational functions to approximate the frequency response. The sampling algorithm is derivative free and well-adapted to devices with rapidly varying frequency responses like microwave filters. The criteria for convergence checking and to determine the location of additional sampling points are easy and fast to evaluate because they are based on the rational functions. Moreover, they provide an estimation of the approximation error and can be used to determine whether the algorithm has problems to reach convergence. The adaptive sampling algorithm leads to a significant reduction of simulation points, as demonstrated by parameter studies. This allows an efficient simulation of electromagnetic responses, as application examples show, which is of great importance when optimizing devices.