# Pastebin RYL5q6ZV
Few of the methods adopted for constraint handling in literature are:
1. choose g_best depending on crowding distance, and use mutation to improve diversity.
2.penalty function, though it has disadvantage for selecting the appropriate factors.
3.By rejection of infeasible solution - but this will decrease diversity.
4. numerical gradient method, I am not sure about this.
5. convert constraint into unconstraint bi-objective- on the basis of crowding distance.
6 . In this method, both feasible and infeasible p_best are stored and updated while only converging feasible towards pareto front and using both feasible and infeasible p_best for deciding g_best. This one I like the best.
or we can work on a completely new idea.