Dear Maurice

Well I do not know what is RSM but comparing the results of GA with ANN depends on the  processed input.

You can check my publications for more information about comparing GA with ANN  i.e.

Multicomponent image segmentation: A comparative analysis between a hybrid genetic algorithm and self-organizing maps

https://www.researchgate.net/publication/261859475_Multicomponent_image_segmentation_A_comparative_analysis_between_a_hybrid_genetic_algorithm_and_self-organizing_maps

Best regards

Article Multicomponent image segmentation: A comparative analysis be...

Sorry Awad. RSM means response surface methodology used to model and optimize a number of input variables for a desired output. I want to compare these two optimization methods RSM and ANN--genetic algorithm.

Hi, you can refer to the following article:

I hope it helps.

http://dl.acm.org/citation.cfm?id=1438768

Hi, if you can be predicted the objective and subjective functions, accurately, then this optimization model can be optimized using a optimization method such as PSO, GWO, HS, GA, ACO, CS, BBO, TLBO or another optimization approaches.

You use GA to optimize the weights in the neural net.

It is possible to use GA to optimize the ANN's connecting weights. GA has also been used to optimize the simple ANN's architecture. Your objective should be minimization of the sum of squares of errors in actual vs. the ANN's prediction.

It would help if you had the trained neural network in your workspace. You can use this trained model as an objective function simply by using the function handle. Say the name of the trained model is 'trainednet'. The objective function can then be defined obj_func = predict(trainednet,x). You can use it as objective function by calling @obj_func