Permutation testing for alternative nonlinear models with application to aging curves of refrigerated vehicles

Testing of model fitting for alternative nonlinear model comparisons within the parametric approach is traditionally a difficult topic due to the complexity of studying the null distribution of test statistics. The nonparametric permutation approach is a flexible method, suitable to be implemented for nonlinear models. In this paper, we introduce a novel algorithm within the nonparametric permutation framework able to perform proper inference on parameters of any specified nonlinear model. The algorithm, although general in its kind, offers a well-rounded approach to make inference via permutation test. Finally, we show the usefulness of the proposed method by applying it for making inference on parameters of a nonlinear aging curve for refrigerated vehicles.