In this paper, the on-line optimization of batch reactors under parame
tric uncertainty is considered. A method is presented that estimates t
he likely economic performance of the on-line optimizer. The method of
orthogonal collocation is employed to convert the differential algebr
aic optimization problem (DAOP) of the dynamic optimization into a non
linear program (NLP) and determine the nominal optimum. Based on the r
esulting NLP, the optimization steps are approximated by neighbouring
extremal problems and the average deviation from the true process opti
mum is determined dependent on the measurement error and the parametri
c uncertainty. A back off from the active path and endpoint constraint
s is determined at each optimization step which ensures the feasible o
peration of the process. The method of the average deviation from opti
mum is developed for time optimal problems. The theory is demonstrated
on an example.