On Benchmarking Numerical Methods for Constrained Global Optimization

Various aspects of benchmarking deterministic and stochastic constrained global optimization methods are considered in this talk. Several techniques for testing numerical global optimization algorithms are described with a particular focus on the recently proposed operational zones https://doi.org/10.1016/j.matcom.2016.05.006 and aggregated operational zones https://www.nature.com/articles/s41598-017-18940-4 . A new generator of constrained test problems is also presented which is based on the GKLS generator https://dl.acm.org/doi/10.1145/962437.962444 widely used in the global optimization literature. It extends the previous generation procedure from the box-constrained case to the case of nonlinear constraints. The user has the possibility to fix the difficulty of tests in an intuitive way by choosing several types of constraints. A detailed information (including the global solution) for each of tests is provided to the user.