Dear all,
I am working on the optimization of the activity chain (set of locations) using a Genetic Algorithm (GA). My fitness function consists of (10) variables [y=x1+x2+x3 .......x10]. Some of the variables should be calling in a dynamic way while others work with an offline database.
Is it right to divide the optimization into two phases, firstly, optimization based on dynamic data using an algorithm (A)? Then, work on an algorithm (B) to optimize the activity chain based on the offline database?
Thank you in advance.
@Ali Wont that be more efficient?
https://www.sciencedirect.com/science/article/pii/S0377042705000774
Hello Ali,
I hope you are doing well.
Yes, it is possible.
Based on your questions A and B, the first thing to come to my mind is to attempt to develop a methodology based on bi-level optimization principles (since you are working with two groups/types of variables). Essentially, you would need to optimize the first group of variables in A and then optimize variables in group B(based on the solutions/values that you have for group A).
Also, an interesting concept to include in your bi-level or dynamic approach is repair heuristics. By applying some repair heuristic rules, you can turn a feasible solution from subproblem A to a feasible solution for subproblem B.
/Dimitris
- Badges
- Science topic