The optimization of the transition time of product changing processes in distillation columns is a Time Optimal Control (TOC) problem. This problem can be solved using a robust but slow Iterative Dynamic Programming (IDP) algorithm or a faster but not robust Sequential Quadratic Programming (SQP) algorithm. The performance of both algorithms is highly influenced by the order of the process model used. A mixed algorithm is proposed as an alternative way to overcome the slowness of the IDP and the poor robustness of the SQP algorithms for this type of problems. This mixed algorithm combines a simplified version of the IDP algorithm, that is use to start the optimization problem, leading to a better convergence region where the SQP algorithm can successfully be started. © 1999 Elsevier Science Ltd.23SUPPL. 1S491S494Hartig, F., Keil, F.J., Large-Scale Spherical Fixed Bed ReactorsModeling and Optimization (1993) Ind. Eng. Chern. Res., 32, pp. 424-437Luus, R., Optimal Control by Dynamic Programming using Systematic Reduction in Grid Size (1990) Int. Journal Control, 51, pp. 995-1013Mendes, A., Schroder, M.J., (1998) Time Optimal Control of Distillation Columns, , CHISA-Congress PragueSchmid, C., Biegler, L.T., Quadratic Programming Methods for Reduced Hessian SQP (1994) Comput. Chem. Engng., 18, pp. 817-832Seferlis, P., Hrymak, A.N., Optimization of distillation units using collocation models (1994) A.I.Ch.E. Journal, 40, pp. 813-825Skogestad, S., Morari, M., Understanding the Dynamic Behavior of Distillation Columns (1988) Ind. Eng. Chern. Res., 27, pp. 1848-1862Schroder, A., (1998), Ph-D Thesis, UNICAMP, Brazil