Description A – For Engineering Managers

HPD Methodology & Software, a Brief Summary

(Slightly modified from an early document by a Xerox engineering manager (C.R.) to his upper management; his words are in quotes.)
(This document, in their lingo, still conveys what resonates with managers even today.)

What IS HPD?

"HPD consists of a methodology and software toolset that guides engineers in describing their system in terms of flow, or propagation, of variances, then applies rigorous mathematical treatments to perform variation analysis and synthesis with all component variability taken into account." (Of course, engineers also have many, many simple problems for which they could apply HPD with great ease; also, users do not see the rigorous mathematics incorporated into the toolset.)

How IS HPD Different from Other Tools/Techniques for Handling Functional Flow of Variation?

"HPD's competitive advantage over existing tools is its advanced mathematical treatment of propagation of stochastic properties (e.g., distributions). This is a very difficult mathematical problem and current techniques rely on linearization, RSS, simplifying assumptions (e.g., all Normal distributions), or Monte Carlo techniques to predict responses. Relative to other techniques, HPD has increased flexibility to handle all distribution types, all types of internal and external noise, interactions or cross term effects, and multiple responses. An additional advantage is that HPD's treatment allows decomposition of a system problem such that stochastic outputs from sub-functions can be input to analyze and optimize higher level functions. Monte Carlo techniques, the most popular alternative, cannot operate in this fashion, making it more difficult to treat more complex or system-level functions."

How Does HPD Support the Current System Design and Electro-Mechanical ((EM) Productivity Projects?

"Core to bringing our System and EM productivity and quality to the next level is more rigorous and coherent treatment and analysis of the flow of variation through a system. As predictive models of process capability at higher levels in a system are developed, not only will better decisions be made prior to major investments in system builds, but fewer builds will be required and the systems allocation and convergence will be significantly improved. Thus HPD is a toolset aimed at enabling such rigorous treatment of flow of variation in a system through advanced mathematics. The tool would directly support the Critical Parameter Management Process by giving delivery teams a means to predict and optimize output performance process capability at different levels. Producing Cpk predictions at all levels will soon be required for passing [Xerox'] Phase 3.2 Exit. Thus we need to provide teams with tools that help them generate accurate predictions prior to the design phase."


Note: The above description refers to only one part of HPD's advantages & capabilities, the full set of which has advanced much, much further since then. In addition, HPD will bring total "literacy" in thinking & working with Variability to all technical employees. And, as importantly, the HPD processes and results would significantly enhance their understanding of their subsystems/systems/problems.