This study presents closed-loop results regarding the production of high-density polyethylene with Ziegler-Natta and Phillips catalysis. The multivariable control algorithm based on the generalized predictive control approach is adopted. Additional features are employed to consider multiple sampling rates and desired trajectory for the output constraints in the cases of grade transition. Results show a significant improvement in the output profiles such as melt index and density. This work provides an extension in the predictive control practice applied to polymerization reactors and presents a new approach for the generalized predictive control algorithm capable to control outputs measured at different rates. (c) 2006 Elsevier Ltd. All rights reserved.