A COMPARISON BETWEEN THE GENETIC ALGORITHM AND THE FUNCTION SPECIFICATION METHODS FOR AN INVERSE THERMAL FIELD PROBLEM

Abstract

In principle, it is possible to apply genetic algorithms (GAs) to the solution of inverse problems in the simulation of manufacturing processes. In this context, an inverse problem can be stated as ‘knowing the desired output of a process, what combination of process parameters are required for its achievement?’. Since the simulation of many processes requires the simulation of thermal, solids and/or fluids problems, the application of GAs to inverse process modelling depends on their ability to solve a wide range of inverse field problems. This paper has two major objectives: (a) to demonstrate the application of a GA to a simple inverse thermal field problem, and (b) to compare its performance against a relatively mature technique for the solution of such problems. The results of this study indicate that, despite the relatively large computational cost of GAs, their accuracy and robustness warrants further investigation of their performance in more demanding applications.