Joe Voelkel
Associate Professor
John D. Hromi Center for Quality and Applied Statistics
College of Engineering
Rochester Institute of Technology
E-mail Address: jgvcqa@rit.edu
Web Address: http://www.isc.rit.edu/~636www/faculty/voelkel.htm
Phone: (716) 475-2231
Fax: (716) 475-5959
Address: John D. Hromi Center for Quality and Applied Statistics
College of Engineering
Rochester Institute of Technology
98 Lomb Memorial Dr.
RIT
Rochester, NY 14623-5604

Title: Minimum-Aberration Split-Plot Designs
Duration: 1 hour seminar
Audience: Statisticians familiar with split-plot and fractional-factorial designs.
Abstract: Split-plot designs are often run as fractional designs, frequently by combining two separate fractional-factorial designs. We show how such a design can be greatly improved by extending the minimum aberration concept. We explore several ways to find optimum designs, present a list of over 150 such designs, and show how such designs can improve the inner-outer-array method sometimes used in robust design. In PowerPoint.
Honorarium: Negotiable. Not expected for a local seminar.

Title: Sequential Experimental Designs for Sensitivity Experiments
Duration: 1 hour seminar
Audience: Statisticians with a background in experimental designs and an idea about Bayesian techniques.
Abstract: We look at experiments with binary responses with one factor that are run sequentially, when the goal is to estimate the setting at which probability of failure is, e.g., 10%. The problem in running such experiments is to sequentially determine reasonable settings. We use a industrial problem to showa Bayesian method to solve this problem optimally (and practically!) We take special care in how to specify practical prior distributions.
Honorarium: Negotiable. Not expected for a local seminar.

Title: Estimating pk for Non-Stable Processes
Duration: 1 hour seminar
Audience: Those involved in quality control; statisticians who want more insight into indices.
Abstract: The pk index is only interpretable when the process being studied is stable (and has a normal distribution), but most processes are not stable. We discuss when it is reasonable to calculate a capability index for a non-stable process, show how the pk index can be naturally generalized to many non-stable processes, provide some examples, and show the superiority of this generalized pk. In PowerPoint/Excel.
Honorarium: Negotiable. Not expected for a local seminar.

Title: Harnessing SPC and ANOVA to Improve Complex Processes
Duration: 1 hour seminar
Audience: Engineers & statisticians who need to study their complex processes for product and process improvement.
Abstract: The most-used passive method of process study uses control charts. But there are more sophisticated techniques, called designed studies and incorporating ANOVA (Analysis of Variance), in which factors in addition to time (e.g. cavity number, operator) can be studied. Several examples will be used to show their power, how they combine graphical and statistical ideas, and how they complement active experimentation.
Honorarium: Negotiable. Not expected for a local seminar.

Title: Design of Experiments
Duration: 4-day shortcourse
Audience: Engineers who need to study many factors for product and process improvement.
Abstract: This course emphasizes two-level designs, the "bread-and-butter" designs in industry that suffice for about 80% of industrial experiments. Participants explore these designs in depth, so they will be able to comfortably study 2 to 15 factors in their experiments. The seminar is highly interactive, and can be taught in two 2-day segments for better learning.
Honorarium: Expected.

Title: Harnessing SPC and ANOVA to Improve Complex Processes
Duration: 4-day shortcourse
Audience: Engineers & statisticians who need to study their complex processes for product and process improvement.
Abstract: We integrate the important idea of control charts with the power of ANOVA so the user can apply both features to the study of complex processes. An example is to find the major sources of variation if you run 10 sets of 4-cavity molding machines over 3 shifts. Numerous examples and exercises reinforce ideas. Can be taught in two 2-day segments. (This is the "how-to version" of the one-hour seminar.)
Honorarium: Expected.