A statistical response surface generator

Small scale special purpose analog simulators are finding many applications as teaching aids in the classroom. One type of simulator design is currently being used in graduate level statistics courses at the University of Delaware. It was developed primarily for teaching Statistical Response Surface Technology.

Statistical Response Surface Technology (SRST) is an effective way of optimizing variables in a system or process. Samples at various operating parameters are taken and analyzed according to one of many Response Surface Design techniques. This statistical sampling process can produce an accurate estimate of a maximum or minimum point, or can give the optimum coefficients of a quadratic model.

The Response Surface Generator (RSG) described in this paper incorporates multivariable function simulators and a random error generator to produce a variety of response surface designs. In general, the student is given a problem based on some process or model which is simulated in the RSG. He can manipulate any of the six input variables and then read the output response. The student may be asked to find the minimum or maximum operating point, the minimum or maximum cost, or the quadratic model for the process using SRST techniques.