Design and manufacturing of molded products are subject to uncontrolled variation (noise) and unknown performance behavior and/or requirements (uncertainty). The validity of current Six Sigma approaches for tolerancing and process optimization for multiple part dimensions is explored. Response surfaces for part weight and two part dimensions are developed as a function of multiple process variables for a rectangular part molded of isotactic polypropylene, i-PP. The process capabilities with respect to dimensional consistency and part weight are assessed using standard practices and Monte Carlo analysis. With respect to tuning of manufacturing processes, multicriteria optimization is necessary to ensure the selection of process set-points resulting in an acceptable likelihood of satisfying multiple dimensional specifications. The Extensive Simplex Method is shown to provide reasonable decision support for process optimization based on a linear process model derived from a main effects design of experiments. With respect to on-line quality control of dimensional consistency, part weight was validated as a good estimator of part dimensions, though requiring validation on an application by application basis. (C) 2004 Wiley Periodicals, Inc.