Transcript Joe Donndelinger

S-Model Perspectives on Aggregation of
Preferences and Decision-Making
Joseph A. Donndelinger
General Motors R&D Center
Warren, MI
The Foundation of the S-Model
The Customer is the Arbiter of Value
Used with permission of H. E. Cook
Overview of the S-Model
Customer-Perceived Value
Drives Market Demand
Increased
Value
Sales Volume
Product Value
Product Specifications Drive
Customer-Perceived Value
Competitive
Advantage
Increased
Value
Improved
Function
Price
Product Spec
(e.g. 0-60 Time, Turning Circle)
Baseline Value
S-Model Aggregation of Value
m
n
i 1
j 1
V  V0  i   Vj
Product
Value
Baseline
Value
“Critical”
Specifications:
Interdependent
“Optional”
Specifications:
Independent
An S-Model Influence Diagram
Profit
Margin
Volume
Value
Price
Cost
Competitors

Product
Subsystem
Subsystem


V
Specifications
Subsystem
A Multi-Criteria Application...
Profit
Margin
Volume
Value
Price
Cost
Product Value is
an aggregation of
Competitors
value ratios for
multiple product
Product
specifications
Subsystem
Subsystem

Subsystem


V
Specifications
…and a Single Criterion Application
Profit
Margin
Volume
Value
Price
Cost
Profit is the single
criterion used in
decision making
Competitors

Product
Subsystem
Subsystem


V
Specifications
Subsystem
Management of Uncertainty
• Sources of Uncertainty
– Evaluation of Product Performance
– Translation of Performance to Value
– Marketplace Changes Over Time
• Competitive Action
• Changes in Customer Preferences
• Exogenous Economic Factors
• Strategy for Uncertainty Management
– Quantify and Propagate Uncertainties
– Computational Efficiency Becomes Important
Top Research Issues
• Development of Rigorous Methods for Aggregation of
Engineering Data, addressing:
–
–
–
–
Propagation of Uncertainty
Combination of Subjective and Objective Information
Ramifications for Decision Quality
Bounds / Limits of hierarchical approaches
• Commonality between Decision Support and Optimization
Systems
– Are common problem formulations feasible and practical?
– How should aspirational (or “stretch”) targets be employed?
• Formalized, bi-directional mapping of decision scenarios
to Decision Support Tools