The Black Box Optimization Competition is the first competition in the continuous optimization field where test problems are truly black boxes to participants. The only information given to the optimizer and participant is the dimension of the problem, bounds on all variables, and a pre-defined budget of black box evaluations.

For this second edition of the challenge, the organizers introduced new tracks for multi-objective optimization and split the single objective track between an expensive track with low budget (low number of function evaluations allowed) and a high-budget track.

Thanks to its state-of-the-art SQP algorithm, Artelys Knitro ranked first on the single objective expensive track and 2nd to 5th in the other tracks.

With this repeated success, Artelys Knitro proves that it is a very robust and versatile solver, perfect for various application, even with black-box functions, typically arising in mechanical engineering, optimal control, or chemical processes.

Extract from the competition website:

Black Box optimization refers to a problem setup in which an optimization algorithm is supposed to optimize (e.g. minimize) an objective function through a so-called black-box interface: the algorithm may query the value f(x) for a point x, but it does not obtain gradient information, and in particular it cannot make any assumptions on the analytic form of f (e.g. being linear or quadratic). We think of such an objective function as being wrapped in a black-box. The goal of optimization is to find an as good as possible value f(x) within a predefined time, often defined by the number of available queries to the black box. Problems of this type regularly appear in practice, e.g., when optimizing parameters of a model that is either in fact hidden in a black box (e.g., a third party software library) or just too complex to be modeled explicitly.