Embed Size (px)
Citation preview
Operations Research in SupplyChain Optimization
M.G. SperanzaUniversity of Brescia, Italy
Department of Management Science & Technology
Athens University of Economics and Business
November 12, 2010
The supply chain
Suppliers factories warehouses customers
Goal of supply chain management
To efficiently integrate suppliers, producers, warehouses to produce and distribute
•the right quantity•at the right time•in the right place
while maximizing the total profit andguaranteeing an appropriate service level
Operations Research
• All strategies, techniques, approaches in supply chain management are focused on:
–Global optimization–Uncertainty management
Goal of supply chain management
Operations Research
Supplyplanning
Productionplanning
Distribution planning Customer
Sequential optimization
Contracts – collaboration – Information technology – Decision Support Systems
Supplyplanning
Production planning
Distribution planning Customer
Global optimization
Global optimization
Operations Research
• Search process of the best strategy for the entire supply chain to:
maximize overall value created
– Supply chain value: difference between what the final product is worth to the customer and the effort the supply chain expends in filling the customer’s request
Global optimization
Operations Research
• The supply chain is complex• The different actors have different goals• The supply chain evolves upon time• The system varies (changes, demand, costs)
Why is it difficult?
A need of sophisticated tools,models, algorithms
Operations Research
Costs
Number of facilities
Inventory
Transportation
Facility costs
5-8
An example of complexity
Conflicting objectives
• Lot Size – Inventory
• Inventory – Transportation
• Lead Time – Transportation
• Product Variety – Inventory
• Cost – Customer Service
Conflicting objectives
• Lot Size – Inventory
• Inventory – Transportation
• Lead Time – Transportation
• Product Variety – Inventory
• Cost – Customer Service
inventory costs
inventory costs
transportation costs
A distribution network
shipment frequency?
Main reason for infrequent (full load) transportation: minimize transportation costs
Main reason for frequent transportation:minimize inventory costs
Conflicting objectives
Minimize the sum of inventory and transportation costs
Problem: Minimize transportation or inventory costs?
The value of integration
AB
C
D
100100
10
100
140
10
100
1 vehiclecapacity 100
customers decide when and how muchsupplier decides routes
frequency of serviceis every 2 days
The value of integration
AB
C
D
100100
10
100
140
10
100
Day 150
50
The value of integration
AB
C
D
100100
10
100
140
10
100
Day 150
50 Cost=340
The value of integration
AB
C
D
100100
10
100
140
10
100
Day 250
50
The value of integration
AB
C
D
100100
10
100
140
10
100
Day 250
50
Cost=360
Cost over two days=700
The value of integration
AB
C
D
100100
10
100
140
10
100
1 vehiclecapacity 100
supplier decides when, how much and routes
The value of integration
AB
C
D
100100
10
100
140
10
100
Day 150
50
Cost=210
The value of integration
AB
C
D
100100
10
100
140
10
100
Day 2
50
50 Cost=210
Cost over two days=420(against 700)
2-21
High Low
Low
High
Responsiveness
Cost
Efficient frontier
success
successfailure
Efficiency and complexity
Intuition fails to generate efficientmanagement policies
In global competition companiesneed to be on the efficient frontier
Need for integration and efficiencymakes decision making more complex
Optimal policy
AB
C
D
100100
10
100
140
10
100
15004000D
20002000C
30003000B
10005000A
demandcapacity
Initial inventory=capacityThe supplier has unlimited availabilityCapacity of vehicles 5000? Periodic policy of minimum
transportation cost
Bell et al (1983), Interfaces
A natural solution
AB
C
D
100100
10
100
10
100
15004000D
20002000C
30003000B
10005000A
demandcapacity
2 vehicles
Repeat the red and the green tours every dayDaily transportation cost: 420
3000
2000
1000
1500
A better solution
AB
C
D
100100
10
100
10
100
15004000D
20002000C
30003000B
10005000A
consumptioncapacity
AB
C
D
100
100
140
Day 1 Day 2
2000
3000
3000
2000
3000
2000
Average dailytransportation cost: 380
An integrated problem
0
1
2
3
production
demand and capacity
costs
An integrated problem
0
1
2
3
When and how much to produce? When and how much to distribute?
Routes?Goal: to minimize total cost
RMI vs VMI
RMI: Retailer-Managed Inventory
VMI: Vendor-Managed Inventory
Each retailer decidesits replenishment policy
The supplier monitors theinventory level of eachretailer and decidesthe replenishment policyof each retailer
Cost of VMI with respect to RMI
VMIOrder-up
general case 71%no fixedtransportation cost 92%small production cost 50%no retailer
inventory cost 67%
VMI vs RMI – the OU policy
Cost of VMI with respect to RMI
VMIOrder-up
general case 71%no fixedtransportation cost 92%small production cost 50%no retailer
inventory cost 67%
VMI vs RMI – the OU policy
Conclusions
• Integrated models are very complex
• Large savings can be achieved
• Appropriate models and algorithms(Operations research) are essential
