There are two key phenomena that affect
multiple-stage supply chains
• Floating bottlenecks
• Stage-spanning bottlenecks
Reference: Graves, Tomlin, “Process Flexibility in Supply
Chains,” 2003.
• Floating bottlenecks
Stage 1
Stage 2
1
1
2
2
3
3
Scenario
Probability
Demand
Product 1 Product 2 Product 3
1
2
0.5
0.5
100
150
50
50
150
100
Expected
Shortfall
Stage 1
Stand
Alone
Shortfall
50
0
25
Stage 2
Stand
Alone
Shortall
0
50
25
Supply
Chain
Shortfall
50
50
50
The expected shortfall is larger than the maximum expected stand-alone
stage shortfall as the bottleneck floats between stages
• Stage-spanning bottlenecks
Stage 1
Stage 2
1
1
2
2
3
3
Demand
Product 1 Product 2 Product 3
150
50
150
Stage 1
Stand
Alone
50
Stage 2
Stand
Alone
50
Supply
Chain
Shortfall
100
These two phenomena are called inefficiencies
Simulation shows that these inefficiencies can significantly affect a
supply chain’s performance
Figure 8
Inefficiencies for Pairs Configuration
60%
40%
Pairs, CI
Pairs, CFI
20%
Pairs, CSI
0%
2
3
4
Numb er of Stages
5
6
A pairs strategy performs much worse than a total flexibility
strategy
Figure 10
Configuration Loss for Pairs Configuration
250%
200%
150%
100%
50%
0%
1
2
3
4
Numb er of Stages
5
6
Chain configurations offer very good
protection against the inefficiencies
Figure 11
Configuration Inefficiency For Chains
60%
50%
40%
h=3 Chain
30%
h=2 Chain, CI
20%
Pairs, CI
10%
0%
2
3
4
Number of Stages
5
6
A chain strategy performs very well
Figure 13
Configuration Loss for Chain Configuration
20%
15%
h=3 Chain
10%
h=2 Chain
5%
The maximum
configuration loss for
0%
1 h=3 was
2 0.34% 3
4
Numb er of Stages
5
6
A chain strategy performs very well
Figure 14
Configuration Loss for Chain Configuration
250%
200%
h=3 Chain
150%
h=2 Chain
100%
Pairs
50%
0%
1
2
3
4
Numb er of Stages
5
6
If the number of stages or the number of
products is very large, then an h=3 chain
strategy may be advisable
Figure 15
Configuration Loss for h=2 Strategy as the
Number of Stages and Products Increase
300%
250%
200%
Num Prod = 10
Num Prod = 20
150%
Num Prod = 30
100%
50%
0%
Num Prod = 40
0
5
10
15
Numb er of Stages
20
25
Flexibility guidelines for single-stage
supply chains
• Try to create chains that encompass as many plants
and products as possible (ideally all plants and
products would be part of one single chain)
• Try to equalize the number of plants (measured in
total units of capacity) to which each product in the
chain is directly connected
• try to equalize the number of products (measured in
total units of expected demand) to which each plant
in the chain is directly connected.
Flexibility guidelines for multiple-stage
supply chains
• The guidelines for single-stage supply chains should be
followed to create a chain structure for each of the
supply chain stages.
• In supply chains with a large number of products or
stages, additional flexibility is advisable, especially for
stages in which the capacity is not much greater than the
expected demand.
• This extra layer of flexibility should again be added in
accordance with the above guidelines to create another
chain structure overlaying the initial chain structure.
The key findings
Multiple-stage supply chains suffer from two types of
inefficiencies that affect performance
BUT
A similar strategy of using chain configurations still
works very well