An electronics assembly firm is planning its production staff needs to make a new modem. It has measured one test worker assembling the unit and observed the following data:

Through unit 2 6 20 25 40 Average time 8.4 5.5 4.2 3.7 3.1 Experience shows that learning curves, which describe the ability of a worker to improve his or her productivity as more an more units are produced, often take the form average time = a 1no. units2b where a and b are empirical constants.

(a) Formulate an unconstrained NLP to fit this form to the given data so as to minimize the sum of squared errors.

(b) Use class optimization software to compute at least a local optimum for your curve fitting model starting with a = 15, b = -0.5.