We present a new heuristics to reduce the computing time of multiple shooting based SQP methods for the numerical solution of optimal control problems with inherent homogeneities in the state variables. We exploit them by a transitional model reduction approach. An approximate model is solved and its solution is used as starting point for the optimization of the detailed model. The method results in a speedup as it reduces the number of expensive SQP iterations. Numerical results for a distillation column model are presented. (c) 2006 Published by Elsevier Ltd.