University of Minnesota – Digital Technology Center
December 2002
Intelligent e-Supply Chain Decision
Support
Norman M. Sadeh
Norman M. Sadeh
e
e
-
-
Supply Chain Management Laboratory
Supply Chain Management Laboratory
School of Computer Science
School of Computer Science
Carnegie Mellon University
Carnegie Mellon University
Copyright
©
2002 Norman Sadeh
Outline


Supply Chain Management: New Context
Supply Chain Management: New Context


Agent
Agent
-
-
Based Collaborative Decision Support

Empirical Results


Dynamic Supply Chain Management Practices
Dynamic Supply Chain Management Practices


Early Results
Early Results


TAC
TAC
’
’
03: A Supply Chain Trading Competition
03: A Supply Chain Trading Competition


Summary and Concluding Remarks
Summary and Concluding Remarks
Copyright
©
2002 Norman Sadeh
Supply Chain Management


Planning and coordinating procurement, production and
Planning and coordinating procurement, production and
distribution activities

-
effective delivery of products to
effective delivery of products to
customers
customers


Extends to product design and configuration
Extends to product design and configuration
Copyright
©
2002 Norman Sadeh
Why is SCM Difficult?


Involves multiple organizations
Involves multiple organizations


Each organization tries to satisfy multiple objectives
Each organization tries to satisfy multiple objectives


Cost, timeliness, quality, market share, etc.
Cost, timeliness, quality, market share, etc.


Each organization operates subject to:
Each organization operates subject to:


©
2002 Norman Sadeh
Historical Perspective
Purchasing Production Sales
Distribution
Materials
Management
Manufacturing
Management
Distribution Customers
Suppliers
Suppliers
Internal
Supply Chain
Customers
Buyers/
Sellers
Buyers/
Sellers
Buyers/
Sellers
Buyers/
Sellers
Inventory
Control
e-Commerce
e-Markets/Exchanges
e-Supply Chains
Functional
Silos

Covisint
, E2open exchanges: more emphasis
, E2open exchanges: more emphasis
on supporting collaboration
on supporting collaboration


Requires richer environments
Requires richer environments
•
•
Multiple attributes
Multiple attributes
–
–
not just price
not just price


Lack of adequate standards
Lack of adequate standards


Lack of adequate decision support tools
Lack of adequate decision support tools


Evaluate a large number of options
Evaluate a large number of options


Peer
-
-
To
To
-
-
Peer/local view vs. more
Peer/local view vs. more
lobal
lobal
view
view


P2P Challenge
P2P Challenge
: Coordinating negotiation across
: Coordinating negotiation across
multiple tiers
multiple tiers


Challenge for the Global View
Challenge for the Global View
:
:


Creating the right incentives for information sharing



Emphasis on Mixed Initiative Decision
Emphasis on Mixed Initiative Decision
Support
Support


Don
Don
’
’
t try to automate everything!
t try to automate everything!
Copyright
©
2002 Norman Sadeh
MASCOT Supply Chain Agent
Bidding & Order Mgmt.
Planning & Scheduling
Procurement
Business
Entity
Enterprise
Level
Mascot
Agent
Tier 1
Suppliers
eMarket

decisions :


Bids & Requests for Quotes
Bids & Requests for Quotes
•
•
including real
including real
-
-
time ATP/CTP
time ATP/CTP


Alternative product/subcomponent designs
Alternative product/subcomponent designs


Make
Make
-
-
or
or
-
-
buy decisions
buy decisions


System
eMarket
Portal
Potential
Business
Partner
Potential
Customer
Bidding/Order Mgmt KS
BOM/Flow Config. KS
Demand Explosion KS
Communication KS
Knowledge Sources
Contexts
External
Systems
Outgoing messages
Incoming Events
Scheduling/CTP KS
RFQ/Procurement KS
Current
working
context
Copyright
©
2002 Norman Sadeh
Main Architectural Features


Blackboard

(
KSs
KSs
)
)


Allows for modular & reusable
Allows for modular & reusable
KSs
KSs


Provides for easy integration with legacy systems
Provides for easy integration with legacy systems


Mixed Initiative Control
Mixed Initiative Control


Customizable user profile
Customizable user profile
Copyright
©
2002 Norman Sadeh
Unresolved Issues


Help keep track of

Associated with KS activations, scripts and goals
Copyright
©
2002 Norman Sadeh
Three Levels of Problem Solving


Knowledge Source Activations
Knowledge Source Activations


e.g. Demand Explosion (RFQ1)
e.g. Demand Explosion (RFQ1)


Scripts
Scripts


e.g. Evaluate (RFQ1)
e.g. Evaluate (RFQ1)
1. Copy_Current_Context
2. Incorporate(RFQ1)
3. Demand_Explosion (RFQ1)
4. Reoptimize_Schedule_with_Net_Demand (RFQ1)
5. Procure_Subcomponents_Net_Demand(RFQ1)
6. etc.
• Goals: Search among multiple options
Copyright
©

Issue
Selected
Unresolved
Issue
Selected
Resolution
Method
Activated
Agenda
Item
Modified
Copy of
Context
U/C = User or Controller
Copyright
©
2002 Norman Sadeh
Status


Customized to support coordination between a
Customized to support coordination between a
machine shop and a tool shop at Raytheon
machine shop and a tool shop at Raytheon


Over 150 machine centers & over 100 people
Over 150 machine centers & over 100 people



Coordination
•orders
•requests for bid
•bid acceptance/rejection
•order cancelation/modification
•products
•bid submissions
•revised delivery dates
Supply chain
entity
Supply chain
entity
Supply chain
entity
Supply chain
entity
Copyright
©
2002 Norman Sadeh
The Coordination Challenge


Generate
Generate
robust yet competitive and cost
robust yet competitive and cost
-
-
effective
effective

uncertainty
uncertainty
are both internal and external
are both internal and external


incoming orders, supplies, internal capacity, etc.
incoming orders, supplies, internal capacity, etc.


Is it possible, through dynamic coordination, to reap
Is it possible, through dynamic coordination, to reap
the benefits of
the benefits of
finite scheduling
finite scheduling
, while offsetting the
, while offsetting the
brittleness
brittleness
of its solutions ?
of its solutions ?
Copyright
©
2002 Norman Sadeh
Real-Time Promising(RTP):
General Considerations


Net Demand


Costs & Priorities: order priorities, late delivery
Costs & Priorities: order priorities, late delivery
penalties, inventory costs, etc.
penalties, inventory costs, etc.


Other Tradeoff: Speed versus
Other Tradeoff: Speed versus
“
“
optimality
optimality
”
”


Assess desirability & decide whether to submit quote
Assess desirability & decide whether to submit quote


Micro
Micro
-
-
Boss RTP module
Boss RTP module
: real
: real
-

To
-
-
Promise
Promise


Selective RTP Validation
Selective RTP Validation
Copyright
©
2002 Norman Sadeh
Profitable-To-Promise


Overall Profit = Total_Revenue
Overall Profit = Total_Revenue
-
-
Total_Costs
Total_Costs


Total_Revenue:
Total_Revenue:
Sum over all orders
Sum over all orders


Total_Costs:

Other variations can be considered


e.g. strategic customers, market share
e.g. strategic customers, market share
considerations
considerations
Copyright
©
2002 Norman Sadeh
Empirical Study:
Multiple RFQ Processing Policies


Response:
Response:


Always bid
Always bid
-
-
no due date negotiation
no due date negotiation


Only submit a bid if overall profit increases
Only submit a bid if overall profit increases





Coordinated finite capacity scheduling
Coordinated finite capacity scheduling
Copyright
©
2002 Norman Sadeh
Empirical Study: Assumptions


A lot
A lot
-
-
for
for
-
-
lot make
lot make
-
-
to
to
-
-
order environment
order environment



schedule
schedule
Copyright
©
2002 Norman Sadeh
Evaluation Criteria


Number of bids refused or rejected
Number of bids refused or rejected


Number of tardy orders
Number of tardy orders


Average utilization of the most utilized resource
Average utilization of the most utilized resource


Average supply chain
Average supply chain
leadtimes
leadtimes


Average due
Average due
-
-

Agent 3
Agent 4
Agent 5
Agent 9
Agent 10
Agent 6
Agent 7
Agent 8
External
customers
Supply chain
Tier3 Tier2 Tier1