Multi-objective optimisation using evolutionary algorithms: its application to HPLC separations

Multi-objective optimisation using evolutionary algorithms (EAs) has been applied for the first time in analytical chemistry and, in particular, in the field of chromatography to develop automated optimisation of gradient separations in reversed phase high-performance liquid chromatography (HPLC). This new approach allows an easy and direct definition of chromatographer goals in the optimisation process and provides not with a single biased optimum solution but with a well-populated Pareto front of nondominated solutions. Thus automated optimisation of chromatography no longer needs the use of chromatographic response functions (CRFs) which were the basis for chromatographic quality evaluation during the last decades. Main steps and new tools in the optimisation process are presented, including new modes of initialisation for the evolutionary algorithm. The separation of 11 pharmaceutical residues of environmental concern has been used as a case study data set to show the practical advantages and working procedures in this new approach.