The Simulated Moving Bed process and its recent extensions called Varicol, PowerFeed and ModiCon are studied, in the case where a small number of columns are used, i.e. from three to five. A multiobjective optimization approach, using genetic algorithms and a detailed model of the multicolumn chromatographic process, is applied to optimize each process separately, and allow for comparison of the different operating modes. The non-standard SMB processes achieve better performance than SMB, due to the availability of more degrees of freedom in the operating conditions of the process, namely the way to carry out asynchronous switches for Varicol, and the different flow rates and feed concentration during the switching interval for PowerFeed and for ModiCon, respectively. We also consider the possibility of combining two non-standard operating modes in a new hybrid process, and evaluate also in this case the possible performance. Finally, a critical assessment of the results obtained and of the potential for practical implementation of the different techniques is reported.