Process optimization using a conditional probability computer
A Conditional Probability Computer is used as part of a process optimizing controller which employs the experimental approach to optimization. This controller does not assume a mathematical model of the process response surface, but periodically moves the process set-point in the direction which will most probably bring the process nearer to its optimal setting. The effects of various controller parameters on its response are studied and a practical solution to the problem of the design of the optimizing experiments is given. Some of its abilities and limitations as a process controller dealing with interacting variables are explored experimentally using an electrical analogue as a process.
Master of Science (M.Sc.)