Methods to find the optimum of an unknown function

Problems involving the search for an optimum condition are often encountered in engineering work. The response for a set of values Of the Independent variables can be determined by experiment, but the real function is usually unknown to the experimenter. The purpose of this study is the development of optimumseeking methods for unimodal functions. The proposed methods use the experimental points to fit a quadratic curve and then find the maximum or minimum of the fitted curve. Practical methods were devised for both deterministic and stochastic type problems. Several other search methods were studied and compared with the proposed methods. Five different test curves were used to best these methods. Results showed that the quadratic curve fitting method converged the fastest.Industrial Engineering, Department o