If you are an ASABE member or if your employer has arranged for access to the full-text, click the underlined title below to view.

Don't have access to the full text? Click here for options.

Most crop simulation models include parameters and coefficients that define responses and interactions with environmental and management conditions. This article presents a conceptual framework for using a genetic algorithm to determine cultivar coefficients of crop simulation models. The model’s multi-dimensional outputs were transformed into a scaled and dimensionless fitness evaluation function based on the L1-metric measure of vector distances between the simulated and observed output. The cultivar coefficients were linearly mapped and encoded into genetic algorithm chromosomes. Empirical evidence comparing the performance of the cultivar coefficients calculated by the genetic algorithm to the existing cultivar coefficients that was calculated by other methods was used to evaluate the performance of this approach. The evidence suggests that the genetic algorithm may be as good as or better than the traditional cultivar coefficient calculation method.