COMPUTATIONAL SURVEY OF UNIVARIATE AND MULTIVARIATE LEARNING-CURVE MODELS

This paper presents a computational survey of the various univariate and multivariate learning curve models that have evolved over the past several years. Discussions are presented to show how the models might be used for cost analysis or productivity assessment in engineering management. A computational experiment comparing a univariate model to a bivariate model is presented. While the bivariate model provides only slightly better fit than the univariate model, it does provide more detailed information about factor interactions and better utilization of available data. The results of the computational experiment can be generalized for the appropriateness of multivariate models.