Some possible basic experiments with meta-heuristics can be:

Changing population size or initialization approaches.

Improving the problem encoding.

Refining the fitness function.

Balancing the exploration and exploitation as per need.

Introducing greedy heuristics and merging in any early, late or any stages of the algorithm.

Lot more!

Hello Sawsan Ali

Actually, it is always dependent on the problem-specific. Some algorithms whether a Heuristic algorithm or a meta-heuristic algorithm are good for some problems rest are for other problems. So, the best algorithms do not exist. What exists are differences between algorithms when applied to a particular problem (such as continuous problems, discrete problems, etc.).

You can modify any algorithm,

1. Parameters (offline or online)

2. Populatuion initialization

3. Local search or Global search

4. Slows-down its convergence speed

6. etc.

I recommend you to read the linked paper.

https://dl.acm.org/doi/pdf/10.1145/937503.937505?casa_token=v9MaesTZ7lcAAAAA:5k0yJx7_fdJb-pquogVlI9kxrbsX-mL0S9bXvqG2HKCKlMGVl1m0u_nwHf7cn1W3XLsbjSTvrUED2Q

Best regards

Thanks for your answer