optimization criteria for my neural network using Genetic Algorithm

hello all! I developed a neural network using Genetic Algorithm(GA). the fitness function of GA will be the neural network performance. I mean the less the error my GA is works well. But the Question arises when I can stop my GA hw i will define the stopping criteria for the same. please give me some suggestions.

Answer 1

[Umang Pandey]

Hi Chandrima,

From what I understand, you can fetch the fitness function values correctly from your neural network and want to determine a stopping criterion for the GA.

When it comes to defining a stopping criterion for a GA, there are several approaches that can be considered, depending on the specific optimization problem at hand. Among these approaches, two popular ones are:

1. Setting a maximum number of generations: This involves specifying a predetermined limit on the number of generations the GA will run before terminating.
2. Defining a convergence criterion: This criterion is based on a tolerance limit, where the GA is terminated when the optimal solution undergoes changes smaller than the specified tolerance value.

These are just a couple of examples, and the choice of stopping criterion depends on the nature of the optimization problem.

For further information refer the documentation:

https://www.mathworks.com/help/gads/genetic-algorithm-options.html#f8335