I just heard of the terminology "black box optimization". I am a little confused about what does it mean! as the name suggests and as I learned is that you are trying to design an algorithm that optimizes an objective function but the algorithm doesn't know (or allowed to use) any prior knowledge about the structure of the function?
So what is not allowed in blackbox optimization:
Using any information derived from the analytical expression to adjust the algorithm?
(So if I know that a given function is multimodal and I know it's global minimum beforehand and I'm using a heuristic algorithm so I'm not allowed to adjust the parameters in a certain way that I know it works for this class of functions. Is this correct?
If this is true, then what is the point of black box optimization?
Hi,
you just don't know the objective function (not given in closed form), but you may evaluate it at given points and from these results you may derive and use any information you want. An example might be an optimization problem where you need to conduct simulations to evaluate your solution.
Regards,
Jörg
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Artificial Intelligence