A Manual Approach to One-Dimensional Minimization

Abstract

An interactive method for manually performing one-dimensional searches is presented. The user computes the function at selected points and the results at each iteration are graphically displayed. As the shape of the function becomes apparent the minimum is readily identified. An experiment was designed to compare the performance of the manual scheme with that of an automated one-dimensional minimization method. Statistical analysis showed that at low accuracies the subjects performed as well as the automated procedure and that there was no significant difference among them. However, at higher accuracies the user’s performance seemed to depend on his natural ability and his effective use of the options for magnifying the display.