Transcript Gonçalves - System Dynamics Society

The Role of Governance in Supply
Chains
Paulo Gonçalves
MIT System Dynamics Group
30 Wadsworth St., E53-358A
Cambridge, MA 02142
Phone 617-258-5585
[email protected]
Albany-MIT Ph.D. Colloquium, MIT System Dynamics Group, April 20 2001.
1
Motivation
• Intel faces enormous challenges in managing its
supply chain
– Must produce the right products at the right time in the
right amount
– In environment of rapid growth, increasingly complex
technology, short product life-cycles, long
manufacturing cycle times and high demand variability
• At the same time, Dell the supply chain leader
can require
– Just-in-time delivery, short windows for order changes
or cancellations (and no penalties)
– High supplier flexibility in product customization
Albany-MIT Ph.D. Colloquium, MIT System Dynamics Group, April 20 2001.
2
The problem
• Traditionally strong supply chain players have
used their leadership position to own advantage
• Self-interested actions can increase own benefits
at the expense of other players
– Manufacturers would like to ensure a steady flow of
orders and maximize volume purchases
– Retailers would like to minimize inventory holding and
obsolescence costs, maintaining quality level
• Locally rational behavior can lead to
inefficiencies
Albany-MIT Ph.D. Colloquium, MIT System Dynamics Group, April 20 2001.
3
Research Questions
• Does it always make sense to act in a narrowly
conceived self-interested way to try to maximize
profits in a supply chain?
• Under what conditions does cooperation and risk
sharing among supply chain players make
sense?
• What cooperative policies in a supply chain are
most appropriate to improve firms’ performance?
Albany-MIT Ph.D. Colloquium, MIT System Dynamics Group, April 20 2001.
4
Purpose and Goals
• To develop a system dynamics model addressing
the issue of governance in a real supply chain
that incorporates several features of real supply
chains often not considered in models in other
literatures, including:
– Explicit behavior rules, instead of myopic and
intertemporal optimization
– Inventory shortages and capacity constraints
– Double ordering dynamics and lost sales dynamics
– Locally available and distorted information
• To develop a set of policies to improve system
performance
Albany-MIT Ph.D. Colloquium, MIT System Dynamics Group, April 20 2001.
5
Relevant Literature
• System dynamics
– Beer game dynamics (Forrester)
– Experimental research (Sterman, Diehl & Sterman,
Croson)
• Microeconomics
– Industrial Organization (Spencer, Williamson, Hart)
– Game theory, incentives and contracts (Tirole)
• Operations management
– Multi-echelon inventory management (Clark &Scarf,
– Supply chain management (Lee at al.,Cachon &
Lariviere)
Albany-MIT Ph.D. Colloquium, MIT System Dynamics Group, April 20 2001.
6
Dynamic Hypothesis
• Narrowly conceived decisions, which are
locally beneficial and boundedly rational,
aimed at maximizing firm performance may,
in a highly complex system, generate
unanticipated side effects that are not in the
best interest of the firm.
Albany-MIT Ph.D. Colloquium, MIT System Dynamics Group, April 20 2001.
7
The Approach
• Develop small concept models for comparative
purposes providing
– Deep understanding of limitations and assumptions of
exiting models in other literatures
– Basis for integrated model and realistic conditions
• Test integrated model in one or two case studies
– PC Industry: Intel - Dell
– Consumer goods industry: P&G - Walmart
Albany-MIT Ph.D. Colloquium, MIT System Dynamics Group, April 20 2001.
8
A microeconomics perspective
Double Marginalization
Total
Profits
Manufacturer
Profits
Manufacturer
Costs
+
+
+
+
Retailer
Profits
+
-
+
Switch
Centralized
Chain
Manufacturer
Revenues
+
Retailer
Costs
Retailer
Revenues
+
+
+
Production
Wholesale
Cost
+
Price
Wholesale
Payment
+
<Reference
Price>
Production
+
Manufacturer
+
Shipments to
Retailer
+
Retail
Price
Retailer
Sales+
Reference
Price
Market +
Demand
+
+
Retailer
Orders
•
•
•
<Wholesale
Price>
Total
Demand
Focus on the financials, complete neglecting the physics
Feedback poor, no dynamics, stationary demand
Unlimited capacity, no delays, perfect information (price and demand),
fully rational behavior, single period maximization
Albany-MIT Ph.D. Colloquium, MIT System Dynamics Group, April 20 2001.
9
An Operations Management Perspective
• Relaxing old assumptions makes models more
realistic, still very complicated. Approach allows:
– Decentralized control
– Multiple decision makers
– Locally rational behavior
• Leading to inefficiencies dealt with contractual
arrangements to improve system performance
–
–
–
–
–
Specifying decision rights: RPM, Quantity fixing
Pricing schemes, minimum purchase
Quantity flexibility, buy-backs
Allocation policies, lead times
Quality
Albany-MIT Ph.D. Colloquium, MIT System Dynamics Group, April 20 2001.
10
Market
Share
+
Firm's Share
Total
Demand
Desired
Backlog
Total
Attractiveness
+
+
Expected
Delivery
Delay
+
Backlog
Adjustment +
Time to
Adjust
Backlog
+
-
+
Competitors
Attractiveness
Table Eff DD
Table for
Attractiveness
Attractiveness
+
-
Delivery Delay
+ -
Reference
Fraction of
Orders Filled
+
Backlog
Target
Delivery
Delay
Firm
Demand
-
Desired
Shipment
Rate
+
Work In
Process
Production
+ Start Rate
Available
Capacity
B2
+
Completion
Rate
Inventory
Table for
Capacity
Utilization
Inventory
Control
+
<Desired
Shipment Rate>
+
Maximum
Shipment
Rate
Manufacturing
Time
Desired
Production +
Start Rate
Inventory
Adjustment
Time
+
Shipments
+
B1
Capacity
Utilization
+
Time to
Perceive
Fraction
Order
Fulfillment
+
-
Perceived
Fraction of
Orders Filled
Order
Fulfillment
Rate
Minimum Order
Processing
Time
Desired
Inventory
Albany-MIT Ph.D. Colloquium, MIT System Dynamics Group, April 20 2001.
11