Differences of learning procedures between GA and RBFN methods. While GA learns the fitness between parameters and time-courses, RBFN learn the fitness of parameters. Therefore, numerical integrations used to evaluate the fitness of calculated time-courses are reduced from trial-and-error to one. The simple GA included in RBFN is used as an input data selection method of RBFN. RBFN enable to fast optimization by reducing the iterative calculations of numerical integrations.