An empirical approach based on the peak separation function has been investigated for use in optimizing chromatographic systems. The approach provides the chromatographer with the ability to specify realistic chromatographic performance goals, i.e., minimum acceptable peak separation and maximum acceptable analysis time. The quality of a separation evaluated in this manner, termed the chromatographic response function (CRF), inherently reflects these performance goals and as such provides a convenient means to order experimental results for use with the simplex optimization algorithm. This scheme has been used to optimize the gradient separation of five PTH-amino acids, involving five experimental variables. Factorial analysis in the region of the optimum indicates that the initial solvent composition is of major importance when only resolution is considered. The gradient parameters provide a means to satisfy both the resolution and the time goal.