Characteristics of optimization procedures

Whatever the optimization procedure may be, it is safe to say that the length of the procedure, i.e. the number of experiments (chromatograms) required, will tend to increase with an increase in the number of parameters involved. Hence, increasing the number of parameters considered leads to 

increased complexity of the response surface, i.e. more local optima, 

This latter advantage needs to be balanced against the three considerable disadvantages. This makes the choice of which parameters to consider during the optimization process (chapter 3) even more important. 

The process may be simplified by defining sensible limits for the parameters. For instance, GC columns have a specified maximum temperature, but for continuous operation a practical maximum well below that value is usually observed. Considering the parameter temperature at a value above the practical maximum is then a waste of time and effort, which may result in optimum conditions that will not be used in practice (e.g. an optimum temperature that equals the specified maximum). 

After the relevant parameters have been selected and their limits have been defined, the actual optimization starts with some initial experiments. A sound a priori prediction of chromatographic retention behaviour is not usually possible for known samples (see chapter 3) and always impossible for unknown samples. Hence, the initial data for the optimization process will have to be provided by some sensibly selected experiments. The initial experiments to be performed will depend on the optimization procedure. They will be discussed in the following sections. 

---