How can I define ANN as an objective function for GA? I am a researcher with no background in programming and I am going to run ANN via toolbox and then use it as an objective function for optimization through GA with Optimization Toolbox.

I am doing a research on finding the best match of AI algorithms for building energy optimization. I have both continues and discrete variables. I am going to use ANN and SVM as objective functions for optimization algorithms such as GA and PSO. I do not have programming background so, I can just use Matlab tool boxes. I was wondering if anyone can help me with this regard.

[Walter Roberson]

What would you like the output of the ANN to be?

[Seyed Saeed Banihashemi Namini]

Thanks Walter. The ANN acts as a prediction model and should be used as an objective function for GA. I am just wondering if I know how I can define the developed ANN through nftool for GA toolbox.

[Raihan Arfin]

Actually I too need the help in the very same regard... I have used ANN for modeling of extraction yield of acid with four input, 3 neurons in hidden layer and one output, the R^2 values are also good around 0.988, but I am finding difficulties in using this for optimization using genetic algorithm... Can anybody please help!!!