The possibilities and limitations of process modelling for online proc
ess and product quality control are described and evaluated. Statistic
al methods and neural networks form the basis of continuous open and c
losed loop quality control, aided by automated adaptive modelling proc
edures. Genetic algorithms support process optimization. The objective
s can be extended to include not only product quality but also process
quality.