A novel nonlinear experimental design scheme to identify the optimum of a multidimensional system was computationally evaluated. The method, a hybrid neural network-genetic algorithm coupled with a fuzzy clustering technique, was developed to accumulate former experimental knowledge and suggest optimal operating conditions for future experimentation. Because of the uncertainty associated with complex systems, several response surfaces with various degrees of complexity and dimensionality were used as case studies for evaluating the new methodology to gain insight into the challenges that may be presented by arbitrary experimental optimization problems in practice. The long-term objective of this work is to develop an experimental design scheme that will discover the best process operating conditions for any difficult process optimization problem in as few experiments as possible. The simulation results demonstrate that the algorithm performance is not significantly affected by parameters such as the number of experiments in a batch of experimentation and convergence criteria for the genetic algorithm as well as for various noise levels below 15%. However, higher noise levels, complexity, and dimensionality decrease its performance. Even at higher levels of noise, complexity, and dimensionality, though, the new approach finds significantly better solutions than those found with traditional statistical experimental design methods, with similar or only slightly more experiments. (c) 2007 American Institute of Chemical Engineers.