Transcript Recent Advances and Future Directions for Quality Engineering

Recent Advances and Future
Directions for Quality Engineering
Geoff Vining
Virginia Tech
USA
Outline
• Recent Advances
– Extending Standard Methodologies to Hard,
Practical Problems
– “Statistical Thinking”
– Applications in Areas Other than Manufacturing
– Advances in Software (However, Need Caution)
– Truly Global Reach
Outline
• Future Directions
– Integrating Quality Engineering Concepts over
Complex Systems
– Large Data Sets
– Dealing with Image Data
– Greater Emphasis on Reliability
– Innovation
– Strong Need to Train Practitioners Properly (Dangers
of Current Software!)
• Statistical Engineering
Background
• Past Department Head of Statistics at Virginia
Tech
• Past Editor of the Journal of Quality
Technology (1998-2000)
• Past Editor of Quality Engineering (2009-2010)
• Past Chair of the ASQ Publications
Management Board
• Member of the ASQ Board of Directors
Background - Journals
• Quality Engineering
– Co-Published by ASQ and Taylor & Francis
– Practitioner Focus
• Journal of Quality Technology
– Published by ASQ
– Focus on High Level Practitioner/Academic
Background - Journals
• Technometrics
– Co-Published by ASQ and ASA
– Similar Focus as JQT, Tends to be More
Mathematical
• Quality and Reliability Engineering
International
– Published by Wiley
– More European
– Publishes “Best” Papers from ENBIS
Extending Standard Methodologies to
Hard Problems
• Experiments with Hard-to-Change and Easyto-Change Factors
– Very Common Practical Problem
– Extensive Literature for Agricultural Applications
– Jones and Nachtsheim
• Profile Monitoring
– Characteristic of Interest Is a Profile (Function)
– Woodall
• Computer Experiments
“Statistical Thinking”
• Originated in the mid 90s
• Basic Idea:
– All work occurs in a system of interconnected
processes.
– Variation exists in all processes.
– The keys to success are:
• understanding variation
• reducing variation.
“Statistical Thinking”
• Roger Hoerl and Ron Snee (2012) Statistical
Thinking: Improving Business Performance
(Wiley and SAS Business Series)
• Point: Biggest contribution that quality
practitioners can make: get senior managers
to understand variation and its sources
• Data Analysis in North America is easy to send
off-shore!
Applications in Areas
Other than Manufacturing
• See Quality Engineering for Examples
• Service Functions
– Accounts Payable
– Product Delivery
– Costumer Relations
•
•
•
•
Risk Management
Security
Healthcare
Several People in Israel Have Done Very Nice
Work!
Advances in Software
• Current Software Can Do Much More
Sophisticated Statistical Analyses to Support
Quality Engineering
– Hard-to-Change versus Easy-to-Change Factors
– Integrated Variance Optimal Designs
– Space-Filling Designs (Computer Experiments)
– Gaussian Stochastic Processes (Comp. Exp.)
Advances in Software
• Exercise Caution with Software “Claims”!
• You Do Not Need to Think about Data
Collection
– Give Us the Factors and the Levels
– We Give You the Plan
• You Do Not Need to Think about The Analysis
– We Plan the Data Collection
– We Know the Best Analysis
• Consequence: Potential for Major Disasters!
Advances in Software
• Software Is an Extremely Important Tool
– Requires Intelligent Use
– “Fisher in a Box”/”George Box in a Box” Does Not
Exist!
• Data Collection Requires Intelligent
Collaboration
– Ask the Right Questions
– Think Carefully about the Science
– Translate Everything Properly into the Analysis
Global Reach
• Foundations to Quality Engineering are North
American and Japanese Manufacturing
– North America: Statistical Theory and Methods
– Japan:
• Quality Management
• “Soft Tools”
• Teamwork
– Deming, Box, Taguchi, “The Gurus”
Global Reach
• Important Influences
– Movement of Manufacturing Away from North
America
•
•
•
•
Asian Tigers
China
India
Latin America (Brazil and Mexico)
– Recognition in Europe of Need for Quality
Engineering: ENBIS
Impact of Global Reach
• Editorial Boards Are Truly Global
• Authors Publishing in the Quality Engineering
Journals Are Truly Global
• Proliferation of Outstanding Quality
Engineering Conferences
• ASQ - Global
Future Directions
• Current Directions Will Continue to Grow
• New Directions
– “Research”
– “Practice”
– Be Aware of the Divide!
Integrating Quality Engineering
Concepts Across Complex Processes
• Complex Processes as Opposed to “Data
Mining” (Next Topic!)
– Developmental Testing of Weapon Systems
•
•
•
•
Multiyear
Multistage
Different Objectives at Each Stage
Competing Interests!
– Complex Manufacturing Processes
• Multistage
• Often, Multi-location
Integrating Quality Engineering
Concepts Across Complex Processes
• Good Quality Engineering Practices May or
May Not Being Used at Substages
• In Some Cases, Just Applying Current Quality
Engineering Methods to the System Work
• In Many More Cases, Need New Methodology
– Formal/Informal Bayesian Methods
– Belief Networks
Large Data Sets
• Data Mining Is Becoming Extremely Important
• Great Deal of Good Work in Israel!
• Emergence of Massive Data Warehouses
(Planet Scale!)
• Standard Statistical Approaches
– Not Valid
– Not Informative
• Often Most Interesting Phenomena: Outliers!
Image Data
• Ability to Monitor Processes via Image Data
• Basic Analysis of Image Data Becoming
“Mature”
• In Some Cases, May Be Able to Adapt
Standard Statistical Process Control
Techniques
• In Many Cases, Must Create New Monitoring
Procedures Based on Image of Every Item
Greater Emphasis on Reliability
• Reliability: Quality Over Time
• Customers Beginning to Demand Highly
Reliable Products and Processes
• Simple Accelerated Life Tests Not Sufficient
• Strong Need:
– Experimental Design and Analysis for Reliabilty
Data
– Process Control with Reliability Data
Innovation
• Not Long Ago, Building Better Quality Was
Significant Innovation
• High Quality Now Viewed as Expectation
• New Issue: Next Way to “Delight” Customers
– “Improved” Current Products
– New Products Customers Never Imagined
• Issue: How Can Quality Engineering Facilitate
Innovation
• See January 2012 Issue of Quality Engineering
Proper Training of Practitioners
• Six Sigma Brought Quality Engineering into the
Hands of Subject Matter Experts
– Typical Training Barely Scratched Surface
– “3 Month Wonders”
– Often, Do Not Know When to Call an Expert
• Software Developments
• Proper Follow-Up Training Essential
Statistical Engineering
• How to best use known statistical principles and tools
to solve high impact problems for the benefit of
humanity.
– tactical integration of statistical thinking with the
application of statistical methods and tools (at the
operational level
– drive proper application of statistical methods based on
solid understanding of statistical thinking principles.
– typically involves the appropriate selection and use of
multiple statistical tools, integrated into a comprehensive
approach to solving complex problems.
• Focus on Large, Unstructured, Complex Problems
• Most Recent Issue of Quality Engineering (April 2012)