It depends on the optimization problem, if its single objective (GA, PSO,...) will work well, if it is scheduling kind (Tabu search...) will do well, If you are working on multi-objective optimization algorithms (NSGA-III, MOPSO) will do well

Why do you limit your choice of methods to heuristics, aren't you interested in optimal solutions for the problems?

So you should try mathematical optimization methods. A lot of MILPs can be solved with current hard- and software in reasonable time. Only if your problem is of a kind that no mathematical formulation of it can be solved in the available timeframe you can resort to heuristics - and in that case it should be tailored for the problem at hand, not just any fancy metaheuristic scheme.

Hi Khairul, you might want to look at Prof. Alireza Soroudi's work. He covers many interesting topics around this subject. Good luck!

What about using normal techniques such as lambda iteration method?

I am not too familiar with this but if memory serves me right, I have seen "lambda iteration" method being compared to simulated annealing and particle swarm optimization in power expansion - I believe it is for economic emission/load dispatch. You might want to search around the internet using these key words. Good luck.

Have a look e.g. at the work of Miguel Carrion to see that power system optimization really is possible using mathematical optimization instead of metaheuristics!

Article Metaheuristic Algorithms in Optimal Power Flow Analysis: A Q...

You can take a look at our recently published review article focusing on optimal power flow analysis for different power system optimization under different metaheuristics