Software for experimental design: the computer program EXCAD

A computer program (EXCAD) dedicated to the optimization of experimental designs to estimate parameters of a mathematical model, is presented. EXCAD computes D-optimal designs and sequentially augmented designs. D-optimal designs minimize the determinant of the variance-covariance matrix and parameters, thus obtaining the average most accurate estimate of parameters. D-optimal designs have generally as many support points as the number of parameters in the mathematical model, so sample scheduling is minimal, not extensive. Augmented designs add to an original design the point that maximizes the decrement of the determinant of the variance-covariance matrix. The general model, linearly or not linearly parametrized Y = F(X,P), that relates two independent variables and P parameters to different responses may be written in the program, while a set of prewritten models is provided.