SSME PARAMETER MODEL INPUT SELECTION USING GENETIC ALGORITHMS

Genetic algorithm are used for the systematic selection of inputs for a parameter modeling system based on a neural network function approxi mator. Due to the nature of the underlying system, issues such as lear ning, generalization, exploitation, and robustness are also examined. In the application considered, modeling critical parameters of the Spa ce Shuttle ;Main Engine (SSME), the Functional relationships among mea sured parameters are unknown and complex. Furthermore, the number of p ossible input parameters is quite large. Many approaches have been pro posed for input selection, but they are either not possible due to ins ufficient instrumentation, are subjective, or they do not consider the complex multivariate relationships between parameters. Due to the opt imization and space searching capabilities of genetic algorithms, they were employed in this study to systematize the input selection proces s. The results suggest that the genetic algorithm can generate paramet er lists of high quality without the explicit use of problem domain kn owledge.