I want to find a minimum of a function on a given interval [a, b] and tried out scipy.optimize.minimize_scalar with method = "Golden" and bracket = [a, b].
The problem is that it returns solutions outside the given interval [a, b]. Is there some way to force the interval or some other scipy/numpy method that can be used?
You can try using the bounded method instead of golden.
method="golden" (i think method="brent" too) take a bracket as a starting guess and may sample outside it.
method="bounded" enforces bounds strictly causing all evaluations and the solution to stay in [a, b]