SEARCH: An Interactive Computer Program for Optimizing Two-Variable, Unconstrained Experiments or Simulation

摘要

The purpose of this study was to develop a computer program that incorporates efficient techniques for solving optimization of experimental test or simulation models. The program is interactive and user-friendly. The program is written in Fortran but can be attach to any simulation model or experiment. The program is limited to two independent variables and one dependent variable. The algorithm of the main program is steepest ascent partan. The study compared several gradient methods and found 2-k factorial the most efficient. The study also concluded that Davies, Swann, and Campey (DSC) - Powell was the most useful line search. The study uses an improvement by Faix to the PARTAN method to eliminate the final linear search. The program is designed to efficiently solve second-order equations and less efficiently higher order equations.