ABSTRACT
A mathematical program is proposed to determine an optimum number of pavement clusters, memberships of the pavement samples to clusters, and associated significant explanatory variables. Simulated annealing and all subsets regression was used to solve the mathematical program. Potential multicollinearity issues were examined and addressed. All possible combinations of the explanatory variables were explored to select the best model specification. Six-cluster models were determined to be the optimum solution for the dataset used in this research. The resultant models were applied to the test data set to examine the prediction accuracy. Normalized root-mean-square error was calculated for each of the resultant models. The associated models were robust with small prediction errors.
