Basin hopping is available in Python via the basinhopping() SciPy function.
The function takes the name of the objective function to be minimized and the initial starting point.
…
perform the basin hopping search
result = basinhopping(objective, pt)
Another important hyperparameter is the number of iterations to run the search set via the “niter” argument and defaults to 100.
This can be set to thousands of iterations or more.
perform the basin hopping search
result = basinhopping(objective, pt, niter=10000)
The amount of perturbation applied to the candidate solution can be controlled via the “stepsize” that defines the maximum amount of change applied in the context of the bounds of the problem domain. By default, this is set to 0.5 but should be set to something reasonable in the domain that might allow the search to find a new basin.
For example, if the reasonable bounds of a search space were -100 to 100, then perhaps a step size of 5.0 or 10.0 units would be appropriate (e.g. 2.5% or 5% of the domain).
perform the basin hopping search
result = basinhopping(objective, pt, stepsize=10.0)
By default, the local search algorithm used is the “L-BFGS-B” algorithm.
This can be changed by setting the “minimizer_kwargs” argument to a directory with a key of “method” and the value as the name of the local search algorithm to use, such as “nelder-mead.” Any of the local search algorithms provided by the SciPy library can be used.
…
perform the basin hopping search
result = basinhopping(objective, pt, minimizer_kwargs={‘method’:‘nelder-mead’})
The result of the search is a OptimizeResult object where properties can be accessed like a dictionary. The success (or not) of the search can be accessed via the ‘success‘ or ‘message‘ key.
The total number of function evaluations can be accessed via ‘nfev‘ and the optimal input found for the search is accessible via the ‘x‘ key.