My question might seem trivial but after much searching I am unable to find the answer.
My question is that is it possible to use gradient descent for finding the minima (conversely maxima as well) while searching a physical space. As an example, we are interested in finding the minima or maxima of a radio signal broadcasted in space. The signal will have a global minima as well as local minima as it propagates in space. The signal is not changing over time. For this radiation pattern is it possible or advisable to implement gradient descent for searching the minima/maxima? If yes than, how would you model the cost function in this case?
Thanks!