Using a tree structured genetic algorithm to perform symbolic regression

A tree structured genetic algorithm is described. The algorithm is used to generate nonlinear models from process input output data. Three examples are utilised to demonstrate the applicability of the technique within the domain of process engineering.