Advantage of multi-objective optimization over single objective What are the advantages of multi-objective optimization over single objective?
I am specifically thinking about MO and SO in Genetic Algorithm.
I have surfed the net and found many articles talking about advantages of multi-objective optimization. Most of them are wordy and make no sense. They just keep writing about optimization of a system with multiple desirable criteria while I do not consider them as advantages of multiple objective optimization.
In multiple objective optimization we find a pareto-optimal solution set. And at the end, we apply weights to make a trade off between the criteria. This is exactly what single objective does from the beginning. So, what is the advantage of multi-objective optimization over single objective optimization.
Myth: Multi-objective optimization is for problem with multiple objectives while single objective optimization is for problems with single objective.
Truth: Both single and multiple objective optimizations can handle problems with multiple objectives.
 A: If you are the decisionmaker yourself and know which objective is more important than others then MO does not give you any advantage over SO.However, I know two situations where MO does make sense:


*

*You are the person that has to solve the problem but are not the decision maker. In this case, you might compute a (small) set of solutions and present all of them to the decision maker who then chooses one of these to implement.

*Solving your problem takes prohibitive long time but the specific instance to solve is not known long before a decision has to be made. In this case, you can input all parameters which are known long enough and treat the remaining parameters as objectives. Then you compute a (possible large) set of solutions upfront and once all remaining parameters are fixed, you only need to search through this set for a best possible solution.
A: Both MOO and SOO may lead to the same result but there are conditions that MOO may be a better tool. If the functions of the optimization problem is not sensible to the decision maker, it is not clear how to chose the weighing factors. in these conditions, MOO performs much better which doesn't require weighing factors.
