Holt et al.'s model was one of the earliest mathematical programming approaches to production planning and control. The model uses quadratic cost functions and the coefficients of the functions are determined by curve fitting techniques from data obtained on previous operation of the plant. Regression-based approaches fit a curve which represents the average trend prevailing in the observed data, whereas, using the technique of Data Envelopment Analysis, it is possible to identify a cost efficient frontier. Estimated coefficients based on such a cost efficient frontier represent the best performance prevailing in the observed data. The two approaches for estimating the coefficients of the cost functions and the corresponding production plans are compared using a numerical example.
