ANN model in this case may be utilized as a fitness function for GA. However, you need to specify how you want to use the hybrid of GA and ANN. A possible scenario is the application of GA for feature selection, which may not be useful to your case. Alternative scenario is to optimize values of independent variables using GA, where you can use the 30 experiments as training data for the ANN, followed by testing on the output population of GA in each generation. ANN error would be the fitness value in this case, and the generation with lowest error would demonstrate the optimized arrays of your 4 factors.

I have attached the matlab code that I have came up with so far. The problem is every time I run my code on the sample data I am getting different results. Also, I am not able to perform any validation checks on the data. Can you please look at the code and suggest what could be going wrong in this case ?

Please read the attached file carefully.

I suspect you have not built in the sampling codes at the initiation stage.

Soumitra K. Mallick

for Soumitra K. Mallick, Nick Hamburger, Sandipan Mallick