How to optimize a neural network using genetic algorithm?

I've a trained NN with 7 input variables X (design parameters) and 1 output variable Y. I want to optimize this surrogate NN model, preferably using the Optimization app. The process should give me the optimized values of the 7 design parameters X, which will lead to minimum Y. Is there a way to do that?

Please advise.

[Sam Chak]

Hi @Bidisha Barman,

Please correct me if I have misunderstood your problem:

The NN model was trained with a fixed dataset consisting of an array of seven (7) inputs, denoted as X, and a corresponding output vector Y.

Your objective is to identify the set of 7 inputs, X, that yield the minimum value of Y from the dataset, regardless of the specific NN model used.

[Harry]

Or, following on from @Sam Chak's comment, is your question how to identify the optimal NN scheme (in terms of number of nodes and layers) given that you have 7 inputs to 1 output?

Answer 1

[Umang Pandey]

Hi Bidisha,

You can refer to the following resource for information on how to optimize a neural network using genetic algorithm:

• This MATLAB answer by MathWorks Support Staff provides insights on how to use Global Optimization Toolbox "ga" to optimize the Neural Network : https://www.mathworks.com/matlabcentral/answers/100323-how-can-i-use-the-genetic-algorithm-ga-to-train-a-neural- network-in-neural-network-toolbox

Hope this helps!