HPLC optimization strategy consists of four elements; experimental design, retention modeling, quality criteria function, and optimum search method. In this paper we present a simple, superior alternative to general classes of classical resolution functions (S function) and a novel optimum search algorithm (iterative stochastic search, ISS) for HPLC optimization. Comparison of S with general classes of resolution-based quality criteria functions (Rs, Rp, and Rmin) shows superior features such as correct assessment of favorable separation conditions, preservation of peak pair contributions, elimination of arbitrary cut-off values, and a unique capability to interpret absolute significance of function values through a simple inequality. The proposed ISS algorithm is more robust than standard methods and it is easily applicable to hyperdimensional optimization. ISS also shows clear advantages in its ability to correctly identify the global optimum (instead of local optimum), with higher precision, with more efficient use of computation cycles, and with easier implementation. Successful application of S and ISS to HPLC optimization was demonstrated in the separation of representative functionalities (sugars, alcohols, and organic acids) present in microbial fermentations. Both the optimal and pathological (worst) conditions were successfully predicted and experimentally verified.