A generalized approach to the use of matrix growth models.

Matrix models provide one of many alternatives to growth and yield monitoring. An attractive aspect of the matrix approach is the ease with which a model may be developed once the appropriate data are collected, but a serious drawback is the inflexibility of matrix models for making growth projections over intervals that are not integer multiples of the remeasurement period. A matrix factorization approach is presented that overcomes this restriction. The key element is the solution of a nonlinear program designed to find an approximation of the original growth matrix.-Authors