MEEN40790 Supply Chain Design &

Analysis

Network Design – Capacitated Facility

Location Model

UCD School of Mechanical and Scoil na hInnealtóireachta

Materials Engineering Meicniúla agus Ábhar UCD
At the end of the session you will know

• Identify the questions to ask about which

facilities to open
• Determine the optimal network design
– Formulate the capacitated facility location model
for a two-stage supply chain
– Apply Excel Solver to determine the optimal
solution
– Extend capacitated facility location model to multi-
stage
• Illustrate supply chain optimisation in action with
industry examples

2
 Recap - Demand Allocation Model

n m
• Which market is served
Min  cij xij
by which plant? i =1 j =1
• Which supply sources are s.t.
used by a plant?
n
xij = volume shipped from
supply site i to customer j
x
i =1
ij
= D j , j = 1,..., m
m

• Assumption
x
j =1
ij
 K i , i = 1,..., n
– Supply chain is in steady
state operation x ij
0
• Supply locations are fixed and
known
• Demand locations are known

3
Supply Chain Design – Blank Page

What questions should a supply chain manager

be asking now? 4
 Global Picture

Tier 1
Suppliers

Manufacturing

Distribution
Centres

Customers
5
 Supply Chain Re-Design

6
What questions should a supply chain manager be asking now?
 Demand Allocation Model Formulation

n m
Min  cij xij • Demand Allocation Model
i =1 j =1 assumes every facility is
open
s.t.
n
• How to modify the model to

x
i =1
ij
= D j , j = 1,..., m capture open or closed
scenarios?
m

x
j =1
ij
 K i , i = 1,..., n

x ij
0

7
 Capacitated Facility Location Model

• Decision variables
– Is factory i open or not
– If factory i is open, shipment quantities from i to j
• Parameters
– Demand at j, capacity at i, shipping cost between i and j,
fixed cost at i
• Objective function
– Minimise total costs (fixed + variable)
• Constraints
• Can only ship from an open factory
• Sum of shipments from open i cannot exceed its capacity Ki

• Sum of shipments from all open i’s must be equal to the

demand at j
• Sum of shipments from each open i must be greater than or
equal to zero (non-negative) 8
 Capacitated Facility Location Model

n n m
• Minimise
Min  f y +  c x ij ij – Fixed cost + variable
i i
i =1 i =1 j =1 cost
s.t.
n

x = D
i =1
ij j
j = 1,..., m • Customer demand is met

x  K y
j =1
ij i i
i = 1,..., n • Volume shipped from an
open plant cannot exceed
its capacity
y {0,1},
i
i = 1,...n, xij  0 • yi is binary (0 or 1)
• xij are non-negative
9
Capacitated Facility Location Model

General inputs required

n = number of factory locations (supply side)
m = number of markets or demand points (customers)
Dj = annual demand from market j
Ki = capacity of factory i
cij = cost of producing and shipping one unit from factory
i to market j (variable shipping cost)
fi = fixed cost to open factory i, even if volume produced
is zero (e.g. annual lease etc)
yi = if factory i is open or not
xij = quantity shipped from open factory i to market j
 Capacitated Facility Location Model

Manufacturing

Distribution
Centres

11
 Capacitated Facility Location Model

Manufacturing

Distribution
Centres

• Model answers two questions

– Question 1
• Which supply locations are open/closed?
12
 Capacitated Facility Location Model

Manufacturing

Distribution
Centres

• Model answers two questions

– Question 2
• Which open supply locations ship to which demand
locations? 13
1.6.1 Capacitated Plant MIP Problem

14
 1.6.1 Problem

1 2 3

Facility

Customer

1 2 3 4 5

15
16
 1.6.1 Solution

2 3

Facility

Customer

1 2 3 4 5

17
1.6.2 SunOil Problem (Ch 5 Chopra & Meindl)
SunOil is a manufacturer of petrochemical products with world wide sales.
You are the VP of supply chain and have several options available. One
possibility is to set up a facility in each region. Advantage is lower
transportation costs and import duties from other regions. Disadvantage is
plants only produce for local demand and may not exploit economies of
scale. Alternative is to consolidate plants in a few regions to improve
economies of scale but increase transportation and import duties. Other
non-quantifiable trade-offs include competitive environment and political
risk.
The table on next slide provides:
• Annual demand per region
• Variable production, inventory, transportation, duties costs of producing
in one region and meeting demand in other regions.
• Fixed costs incurred independent of volume.
• Variable costs linked to volume produced/shipped.
SunOil is considering two facility sizes, low capacity plants which produce
10,000,000 per year and high capacity producing 20,000,000 per year. High
capacity plants exhibit some economies of scale with fixed costs less than
double those of low capacity plants. What network design should the supply 18

chain VP pursue?
1.6.2 SunOil Problem

Cost of shipping from

N. America to N. America in
thousands $

E.g., Annual Demand Cost for Cost for

for N. America opening a opening a
low-capacity high-capacity
plant plant

19
1.6.2 SunOil Problem

20
 1.6.2 SunOil Problem
North South Europe Africa
Asia
America America

K = 20x106 K = 20x106 K = 20x106 K = 20x106 K = 20x106

K = 10x106 K = 10x106 K = 10x106 K = 10x106 K = 10x106

North South
Europe Asia Africa
America America 21

D = 12x106 D = 8x106 D = 14x106 D= 6x106 D = 7x106

1.6.2 SunOil Problem
North South Europe Asia Africa
America America

K = 20 K = 20 K = 20 K = 20 K = 20

K = 10 K = 10 K = 10 K = 10 K = 10

North South Europe Asia Africa

America America
D = 12 D=8 D = 14 D = 16 D=7
What do you think the optimal network
could look like?
22
1.6.2 SunOil Problem
North South Europe Asia Africa
America America

K = 20 K = 20 K = 20 K = 20 K = 20

K = 10 K = 10 K = 10 K = 10 K = 10

North South Europe Asia Africa

America America 23
D = 12 D=8 D = 14 D = 16 D=7
 1.6.2 Sunoil Solution
Template

Solution

24
1.6.2 SunOil Solution

25
1.6.2 SunOil Solution

North South Europe Asia Africa

America America

K = 20 K = 20 K = 20

North South Europe Asia Africa

America America
D = 12 D=8 D = 14 D = 16 D=7 26
1.6.2 SunOil Optimum Network

What could make the model

more complex? 27
 Model Extension: Accounting for Taxes,
Tariffs and Customer Requirements

• Networks should be structured to maximize profit after

taxes while meeting customer service requirements
• Objective function maximizes profits
m n n n m
Max  r j  xij −  Fi y i − cij xij
j =1 i =1 i =1 i =1 j =1

rj is revenue from selling one unit in market j

• Not all customer demand has to be met

n

xi=1
ij  D j for j = 1,K ,m

Or could put lower and upper bound on customer shipment

 Supply Chain Optimisation in Action

J. David Allen, Roger L. Tobin, Anthony Calderan (2017) Verizon Optimizes Work Center Locations to Reduce Installation and
Repair Operations Costs. Interfaces, Published online in Articles in Advance 10 Mar 2017
http://dx.doi.org/10.1287/inte.2016.0871
 Value of Supply Chain Optimisation –
Pfizer & Wyeth

“Pfizer is continuing its strategy of manufacturing-facility rationalization

and supply-chain optimization as it prepares for the closure and eventual
integration of Wyeth. Pfizer plans to further reduce its global internal
network of plants to 41”

"We have a complex network of manufacturing plants, with excess capacity

that is not good for costs," Nat Ricciardi, Pfizer's president of
manufacturing, said in an interview.

The global job cuts comes in the wake of last year's purchase of rival
Wyeth. Pfizer had 40 manufacturing sites before acquiring more than three
30
dozen Wyeth facilities in the October merger.
 Value of Supply Chain Optimisation:
Merger and Acquisitions─Miller and Coors, then
SABMiller & InBev (2016)

• Miller and Coors merged operations in the U.S. to become

MillerCoors in the summer of 2008.
• They promised investors $750 million in savings.
• They frequently reported on the supply chain network
modeling results:
– “Supply chain integration continues to proceed on schedule., as product
moves are more than 90 percent complete. The brewery optimization
project is nearing completion.”
– “In the third quarter, MillerCoors successfully completed initial product
transitions within its national brewery network. The company will
continue to focus on further network optimization through peak/non-
peak season sourcing changes, as well as opportunities for increased
efficiencies.”

31
Session Summary

• Identify the questions to ask about the design/re-

design of a supply chain network
• Determine the optimal network design
– Formulate the capacitated facility location model
for a two-stage supply chain
– Apply Excel Solver to determine the optimal
solution
– Extend capacitated facility location model to multi-
stage
• Illustrate supply chain optimisation in action with
industry examples

32