Dear sir

I do not have review papers. But I have many papers about optimization. You can see it in my page in researchgate. i hope it it useful for you.

It certainly appears to me that you first need to worry about the properties of the optimization model itself. If you can parameterize accurately the decision variables as well as represent, or calculate, one design's performance, then you can certainly optimize your design MUCH better WITHOUT the use of the too popular but untrustworthy metaheuristics. I would - every day - perform classical nonlinear optimization much rather than relying on a metaheuristic, that can give you bogus results, and probably the metaheuristics would require much more computer time as well. (Besides the fact that you have much less a clue where an optimum might be.)

thanks everyone :-)