In MATLAB there is a inbuilt optimization method called GA. How we can put convergence criteria so that my optimization problem will be terminated. Please see attached figure. In that figure after 100 generation , solution is terminated but result is not converged as we are seeing best and mean value are very far
It depends on the problem at hand. The bigger a better possibility to reach the optimal value
* fitness function value should be less than the given threshold value.
* Difference between best and average fitness is below than a given fraction of the fitness of the average individual.
* Difference between the best individual of the current population and the best individual so far must be very small.
* Number of iterations (sample population) exceeds a given limit.
To set the convergence criteria, evaluate your fitness function in terms of iterations, and based on that you can terminate your algorithm. For example, you can terminate your algorithm if the fitness function value is not improved for specific number of iterations say 100.
you can see the experimental setup of the following paper.
Article CC-GA: A Clustering Coefficient based Genetic Algorithm for ...
- Badges
- Science topic