About Us
Grad Studies
Jump To Topic...
Theme Topic
Black-box Optimization

Black-box optimization problems commonly arise in process systems engineering due to the heavy use of process simulators. Although many simulators use equation-based models derived from theory, the solution of these equations often involves techniques such as sequential modular flowsheeting, table look-ups, the solution of sub-problems (such as minimization of Gibbs free energy), or calls to third-party software for which the underlying equations are hidden from the end user. As a result, most process simulations are effectively black-box.

Optimization of black box problems has been an ongoing area for research, resulting in the development of methods such as moving simplex, particle swarm optimization, differential evolution, genetic algorithms, and the like. Most of these black box solvers are inspired by biology but do not take advantage of the many gains made in white-box optimization theory (mathematical programming, MINLP, etc). Therefore, the focus of this research is to develop new black-box optimization algorithms using the theory and strategies of white-box optimization problems.

Dr. Thomas A. Adams II
Associate Professor