EVALUATING NETWORK DESIGN

DECISIONS USING DECISION

TREES

USING DECISION TREE TO EVALUATE NETWORK DESIGN

AAglass
glassfactory
factoryspecializing
specializingin
incrystal
crystalis
isexperiencing
experiencingaa
substantial
substantialbacklog,
backlog,and
andthe
thefirm's
firm'smanagement
managementis
is
considering
consideringthree
threecourses
coursesof
ofaction:
action:
A)
A) Arrange
Arrangefor
forsubcontracting
subcontracting
B)
B) Construct
Constructnew
new facilities
facilities
C)
C) Do
Donothing
nothing(no
(nochange)
change)
The
Thecorrect
correctchoice
choicedepends
dependslargely
largelyupon
upondemand,
demand,which
which
may
maybe
below,
low,medium,
medium,or
orhigh.
high. By
Byconsensus,
consensus,management
management
estimates
estimatesthe
therespective
respectivedemand
demandprobabilities
probabilitiesas
as0.1,
0.1,0.5,
0.5,
and
and0.4.
0.4.

Example of a Decision Tree Problem (Continued): The Payoff

Table
The
The management
management also
also estimates
estimates the
the profits
profits
when
when choosing
choosing from
from the
the three
three alternatives
alternatives (A,
(A,
B,
B, and
and C)
C) under
under the
the differing
differing probable
probable levels
levels of
of
demand.
demand. These
These profits,
profits, in
in thousands
thousands of
of dollars
dollars
are
are presented
presented in
in the
the table
table below:
below:

A
B
C

0.1
Low
10
-120
20

0.5
Medium
50
25
40

0.4
High
90
200
60

Example of a Decision Tree Problem (Continued): Step 1. We start by

drawing the three decisions

A
B
C

Example of Decision Tree Problem (Continued): Step 2. Add

our possible states of nature, probabilities, and payoffs
High demand (0.4)
Medium demand (0.5)
Low demand (0.1)

High demand (0.4)

Medium demand (0.5)

Low demand (0.1)

High demand (0.4)

Medium demand (0.5)
Low demand (0.1)

$90k
$50k
$10k
$200k
$25k
-$120k
$60k
$40k
$20k

Example of Decision Tree Problem (Continued): Step 3. Determine the

expected value of each decision

$62k
$62k
AA

High
Highdemand
demand(0.4)
(0.4)
Medium
Mediumdemand
demand(0.5)
(0.5)

$90k
$90k
$50k
$50k

Low
Lowdemand
demand(0.1)
(0.1)

$10k
$10k

EV
EVAA=0.4(90)+0.5(50)+0.1(10)=$62k
=0.4(90)+0.5(50)+0.1(10)=$62k

Example of Decision Tree Problem (Continued): Step 4. Make

decision
High demand (0.4)

$62k
A
B

$80.5k

Medium demand (0.5)

Low demand (0.1)
High demand (0.4)
Medium demand (0.5)
Low demand (0.1)

High demand (0.4)

$46k

Medium demand (0.5)

Low demand (0.1)

$90k
$50k
$10k
$200k
$25k
-$120k
$60k
$40k
$20k

Alternative
AlternativeBBgenerates
generatesthe
thegreatest
greatestexpected
expectedprofit,
profit,so
so
our
ourchoice
choiceis
isBBor
orto
toconstruct
constructaanew
newfacility
facility

LATIHAN
Tika adalah perusahaan jasa pengiriman obat-obatan. Untuk lima tahun kedepan,
pemilik perusahaan telah mempertimbangkan untuk memilih satu dari tiga skenario
strategi pengembangan kapasitas berikut ini
Skenario pertama adalah skenario yang paling pesimis yaitu perusahaan tidak
melakukan apapun selama lima tahun kedepan.
Skenario kedua, perusahaan akan meningkatkan kapasitasnya dengan memperbanyak
jenis jasa pengiriman barang yang ditawarkannya (tidak hanya obat-obatan)
Skenario ketiga, perusahaan mengefisiensikan kapasitas yang dimilikinya dengan
memindahkan lokasi usahanya ketempat baru yang menurut perusahaan lebih dekat
dengan sejumlah perusahan yang merupakan langganan sehingga perusahaan dapat
menghemat biaya pengiriman barang.

Bantulah pemilik perusahaan untuk memilih satu skenario pengembangan kapasitas yang
terbaik, apabila:
Perusahaan memilki peluang sebesar 55 persen untuk mendapatkan peningkatan yang sangat
kuat atas permintaan jasa pengiriman barang .
Apabila perusahaan tidak melakukan apapun, peningkatan yang saat kuat atas permintaan jasa
pengiriman barang yang ditawarkan oleh perusahaan tersebut akan memberikan keuntungan
sebesar Rp.55.000.000 pertahun; sedangkan peningkatan permintaan yang rendah hanya akan
memberikan keuntungan sebesar Rp.27.500.000 per tahun.
Apabila perusahaan memperbanyak jenis jasa pengiriman barang yang ditawarkannya (tidak
hanya pengiriman obat-obatan), peningkatan yang sangat kuat atas jasa pengiriman barang
yang ditawarkan oleh perusahaan akan memberikan keuntungan sebesar Rp 107.000.000
pertahun dan peningkatan yang rendah atas jasa pengiriman barang yang ditawarkan oleh
perusahaan hanya akan memberikan keuntungan sebesar Rp. 69.000.000 pertahun. Untuk
memperbanyak jenis jasa pengiriman barang yang ditawarkannya, perusahaan harus
mengeluarkan biaya sebesar Rp. 65.000.000 sebagai investasi awal .
Apabila perusahaan memindahkan lokasi usahanya ke tempat baru, peningkatan yang sangat
kuat atas permintaan jasa pengiriman obat-obatan akan memberikan keuntungan sebesar
Rp.150.000.000 pertahun; sedangkan permintaan yang rendah hanya akan memberikan
keuntungan sebesar Rp. 85.000.000 pertahun. Untuk pendirian kantor baru ini, perusahaan
harus mengeluarkan biaya investasi sebesar Rp.100.000.000 rupiah.

STRATEGIC TIMING: THE

OPTION VALUE OF WAITING
Strategic capacity timing often deals with the essensial question: should we
adjust capacity now or should we wait?
The time frame considered is long (years) and the decision is complicated by
vast uncertainty and significant factors
The timing decision should be based on both strategic and financial factors
Definition of the option of waiting: the value maximization principle favors
the strategy with highest expected net present value we shoudl
calculate the expected NPV of waiting and of immediate expansion, and
then pick the strategy with highest NPV

STRATEGIC TIMING: THE

OPTION VALUE OF WAITING
The strategy selection rule can be rephrased in terms of the option
value of waiting, which is the difference between the expected NPV
of waiting and the expected NPV of expanding
Option value of waiting: NPV (wait)- NPV (do not wait)
The strategy selection rule thus simply becomes wait if and only if it
has a positive option value.

STRATEGIC TIMING: THE OPTION

VALUE OF WAITING- ACER EXPANSION
Acer, a Taiwanese computer company, started opertions by serving the lower-price
segment in Asia. Later, in its quest for a global expansion strategy, it decided to expand
into emerging markets by serving other low-price segment in Mexico, South Africa, and
Russia
The question is wether Acer should expand now, or wether it should wait an
additional year before expanding into emerging market
Assume that the demad forecast for those theree scearios, is respectively: 200 thousand
unit per year with likelihood of 25%, 100 thousand units per year with likelihood of 50%,
or 30 thousand units per year with likeliihood of 25%
The cost structure: Capacity expansion incurs a fixed cost of $ 8 million plus a marginal
cost of $ 50 per unit of capacity
Cost of adding capacity 100.000 units per year???
Accer expect each PC to contribute about $80 in operating profits

STRATEGIC TIMING: THE OPTION VALUE

OF WAITING- ACER EXPANSION

NPV of
expand
now

If Acer decide to expand now, the optimal size

of the capacity addition is 100 thousand units
per year at a capital expenditure of $ 13million

Except for medium demand scenario, the capacity decision is mismatch between supply and
demand

STRATEGIC TIMING: THE OPTION VALUE

OF WAITING- ACER EXPANSION

NPV of
expand
now

Ignoring shortage costs, the operating profit in these

three scenario are $ 8 million, $8 million and $2,4
million 25%x 8+50% x 8+25% x 2,4= $6,6 million
NPV = -13+(6,6/1,25)+ (6,6/1,252)+(6,6/1,253)+
(6,6/1,254) =$2,6 million

STRATEGIC TIMING: THE OPTION VALUE

OF WAITING- ACER EXPANSION

NPV
of
Wait

If acer decides to wait, it can observe its commercial success

in its current home market, which is highly correlated with its
success in the emerging markets
Waiting thus gives accer much better information of demand
Assuming for simplicity a perfect correlation, it can then make
capacity decisions under certainty next year
If demand high, the capacity expansion will add 200 thousand
units ($18 million)
At medium demand, capacity addition is 100 thousand units
per year at $13 million
If demand low, not invest

STRATEGIC TIMING: THE OPTION

VALUE OF WAITING- ACER EXPANSION

The capital expenditure is

(25%x $18)+(50% x $13)+(25%x 0)= $ 11million
The operating profits
(25% x $16)+(50%x$8)+(25%x 0)= $ 8 million
The net present value=
NPV wait= -(11/1,25)+(8/1,252)+(8/1,253)+(8/1,254)=$3,7 million

Latihan
PT Insan Merdeka, Tbk (sebuah perusahaan yang memproduksi alat-alat berat di Indonesia) mulai
mempertimbangkan untuk memasarkan produknya secara global ke negara-negara di kawasan Asia
Tenggara. Saat ini, PT Insan Merdeka, Tbk telah mulai melakukan penjualan produknya ke Negara
Malaysia dan untuk selanjutnya, PT Insan Merdeka, Tbk akan melakukan melakukan penjualan
produknya ke Negara Thailand dan Vietnam.
Bantulah jajaran Direksi dari PT Indan Merdeka, Tbk untuk memutuskan apakah PT. Insan Merdeka, Tbk
harus melakukan penjualan produknya ke Negara Thailand dan Vietnam saat ini atau menunggu sampai
dua tahun yang akan datang. Pertimbangan menunggu dua tahun yang akan datang adalah PT. Insan
Merdeka, Tbk dapat mengamati terlebih dahulu perkembangan permintaan at-alat berat di Malaysia;
dengan asumsi, pola permintaan produk di Malaysia akan sama dengan pola permintaan produk di
Thailand dan Vietnam. Asumsi lain dan data-data lain yang diperlukan dapat diuraikan sebagai berikut:
Terdapat tiga skenario terkait dengan permintaan alat-alat berat di masa yang akan datang: (i)
permintaan alat-alat berat sebesar 400 ribu unit pertahun memiliki peluang untuk terjadi 30%
(tinggi); (ii) permintaan alat-alat berat sebesar 300 ribu unit pertahun memiliki peluang untuk terjadi
50% (moderat); dan (iii) permintaan alat-alat berat sebesar 200 ribu unit pertahun memiliki peluang
untuk terjdi 20% (rendah).

 Cost structure: (i) Biaya tetap: setiap penambahan kapasitas sebesar 100 unit
pertahun membutuhkan biaya tetap sebesar Rp. 8.000.000.000; (ii) biaya variabel:
Rp. 50.000 per unit
Jika penambahan kapasitas dalam rangka penjualan produk ke Thailand dan
Vietnam dilakukan tahun ini maka maka perusahaan hanya akan menambah
kapasitas sebesar 100 ribu unit saja; namun jika penambahan kapasitas dalam
rangka penjualan produk ke Thailand dan Vietnam dilakukan dua tahun yang aka
datang maka penambahan kapasitas akan dilakukan sesuai dengan peramalan dari
tingkat permintaan.
Penjualan produk akan memberikan profit sebesar Rp. 60.000 per unit
Analisa dilakukan untuk periode waktu lima tahun

EVALUATING NETWORK DESIGN DECISIONS

USING DECISION TREES
A manager makes several different decisions when designing a
supply chain network.
If uncertainty is ignored, a manager will always sign long-term
contract (because they are typically cheaper) and avoid all flexible
capacity (because it is more expensive)
Such decisions, however, can hurt the firm if future demand or
prices are not as forecast at the time of the decision

EVALUATING NETWORK DESIGN

DECISIONSUSING DECISION TREES
During network design, managers thus need a methodology that allows them to
estimate the uncertainty in their forecast of demand and price and then
incorporate this uncertainty in the decision-making process.
Such a methodology is most important for network design decisions because
these decisions are hard to change in the short term A decision tree is a
graphic device used to evaluate decisions under uncertainty; decision trees with
DCFs can be used to evaluate supply chain design decisions given uncertainty
in prices, demand, exchange rates, and inflation.

EVALUATING NETWORK DESIGN

DECISIONSUSING DECISION TREES
1.

Identify the duration of each period (month, quarter, etc.) and the number of periods T
over which the decision is to be evaluated.

2.

Identify factors such as demand, price, and exchange rate whose fluctuation will
be considered over the next T periods.

3.

Identify representations of uncertainty for each factor; that is, determine what
distribution to use to model the uncertainty.

4.

Identify the periodic discount rate k for each period.

5.

Represent the decision tree with defined states in each period as well as the
transition probabilities between states in successive periods.

6.

Starting at period T, work back to Period 0 identifying the optimal decision and
the expected cash flows at each step. Expected cash flows at each state in a given
period should be discounted back when included in the previous period.

EVALUATING NETWORK DESIGN DECISIONS

USING DECISION TREES- Trips Logistics
The manager must decide whether to lease warehouse space for the coming three
years and the quantity to lease.
The long term lease is currently cheaper than the spot market rate for warehouse
space. The manager anticipates uncertainty in demand and spot prices for warehouse
space over the corning three years. The long-term lease is cheaper but could go
unused if demand is lower than anticipated. The long-term lease may also end up
being more expensive if future spot market prices come down.
In contrast, spot market rates are high and warehouse space from the spot market will
be expensive if future demand is high. The manager is considering three options:
1. Get all warehousing space from the spot market as needed.
2. Sign a three-year lease for a fixed amount of warehouse space and get additional
requirements from the spot market.
3. Sign a flexible lease with a minimum charge that allows variable usage of warehouse space
up to a limit with additional requirement from the spot market.

EVALUATING NETWORK DESIGN DECISIONS USING DECISION

TREES- Trips Logistics

One thousand square feet of warehouse space is required for every 1,000 units of demand, and the
current demand at Trips Logistics is for 100,000 units per year.
The manager decides to use a multiplicative binomial representation of uncertainty for demand and
price.
From one year to the next, demand may go up by 20 percent with a probability of 0.5 or go down by
20 percent with a probability of 0.5. The probabilities of the two outcomes are unchanged from one
year to the next.
Warehouse space is currently available on the spot market for $1.20 per square foot per year. From
one year to the next, spot prices for warehouse space may go up by 10 percent with probability 0.5
or go down by 10 percent with probability 0.5 according to a binomial process. The probabilities of
the two outcomes are unchanged from one year to the next.
Or The general manager can sign a three-year lease at a price of $1 per square foot per year
Each unit Trips Logistics handles results in revenue of $1.22
Trips Logistics uses a discount rate of k = 0.1 for each of the three years.

EVALUATING NETWORK DESIGN DECISIONS USING DECISION TREES- Trips

Logistics

EVALUATING NETWORK DESIGN DECISIONS USING

DECISION TREES- Trips Logistics

EVALUATING NETWORK DESIGN DECISIONS USING DECISION TREESTrips Logistics

Deman
d

price

Revenue

Cost (D=, p=)

Profit (D=,
p=)

144

1,45

175.680

208.800

(33.120)

144

1,19

175.680

171.360

4.320

144

0,97

175.680

139.680

36.000

96

1,45

117.120

139.200

(22.080)

96

1,19

117.120

114.240

2.880

96

0,97

117.120

93.120

24.000

64

1,45

78.080

92.800

(14.720)

64

1,19

78.080

76.160

1.920

64

0,97

78.080

62.080

16.000

Table 6-5

EP(D = 120, p = 1.32, 1) = 0.25 X [P(D = 144, p = 1.45, 2) + P(D =

144,p = 1.19, 2) + P(D = 96, p = 1.45, 2) + P(D = 96, p = 1.19, 2)] =
0.25 X ( -33,120 + 4,320 - 22,080 + 2,880] = -$12,000

EVALUATING NETWORK DESIGN DECISIONS USING

DECISION TREES- Trips Logistics
The present value of this expected
value in Period 1 is given by
PVEP(D = 120, p = 1.32, 1) = EP(D =
120, p = 1.32, 1)/(1 + k) = -12,000)1.1 = $10,909

The manager obtains the total expected profit

P(D = 120, p = 1.32, 1) at node D = 120, p =
1.32 in Period 1 to be the sum of the profit in
Period 1 at this node and the present value of
future expected profits.

P(D = 120, p = 1.32, 1) = 120,000

X 1.22 - 120,000 X 1.32 + PVEP(D
= 120,p = 1.32, 1) = -$12,000 $10,909 = -$22,909

EVALUATING NETWORK DESIGN DECISIONS USING DECISION TREES- Trips Logistics

The total expected profit for all other nodes in Period 1

3. EVALUATING NETWORK DESIGN DECISIONSUSING

DECISION TREE
For Period 0, the total profit P(D = 100, p = 1.20, 0) is the sum of the profit at Period 0 and
the present value of the expected profit over the four nodes in Period 1.
EP(D = 100, p = 1.20, 0) = 0.25 X [P(D = 120, p = 1.32, 1) + P(D = 120, p = 1.08, 1) + P(D
= 96,p = 1.32, 1) + P(D = 96,p = 1.08, 1)] = 0.25 x_ [ -22,909 + 32,073 - 15,273 + 21,382]
= $3,818
PVEP(D = 100,p = 1.20, 1) = EP(D = 100,p = 1.20, 0)/ (1 + k) = 3,818/1.1 = $3,471
P(D = 100,p = 1.20, 0) = 100,000 X 1.22 - 100,000 X 1.20 + PVEP(D = 100,p = 1.20, 0) =
$2,000 + $3,471 = $5,471

EVALUATING NETWORK DESIGN DECISIONS USING

DECISION TREES- Trips Logistics
The manager next evaluates the alternative where the lease for 100,000 sq. ft. of warehouse
space is signed.
The evaluation procedure is very similar to the previous case, but the outcome in terms of profit
changes.
For example, at the node D = 144,p = 1.45, the manager will require 44,000 sq. ft. of warehouse
space from the spot market at $1.45 per square foot because only 100,000 sq. ft. has been leased
at $1 per square foot.
If demand happens to be less than 100,000 units, Trips Logistics still has to pay for the entire
100,000 sq. ft. of leased space.

EVALUATING NETWORK DESIGN DECISIONS USING

DECISION TREES- Trips Logistics
The manager next evaluates the total expected profit for each node in Period 1.

EVALUATING NETWORK DESIGN DECISIONS USING

DECISION TREES- Trips Logistics
For Period 0, the expected profit EP(D = 100, p = 1.20, 0) over the four nodes in Period 1 is
given by:
EP(D = 100,p = 1.20, 0) = 0.25 X [P(D = 120,p = 1.32, 1) + P(D = 120, p = 1.08, 1) + P(D = 96,p
= 1.32, 1) + P(D = 96,p = 1.08, 1)] = 0.25 X [35,782 + 45,382 - 4,582 - 4,582] = $18,000
The present value of the expected profit in Period 0 is given by: PVEP(D = 100,p = 1.20, 0) =
EP(D = 100,p = 1.20, 0)/(1 + k)= 18,000/1.1 = $16,364
The total expected profit is obtained as the sum of the profit in Period 0 and the present value of
the expected profit over all four nodes in Period l.lt is
P(D = 100,p = 1.20, 0) = 100,000 X 1.22 - 100,000 X 1 + PVEP(D = 100, p = 1.20, 0) = $22,000
+ $16,364 = $38,364
The NPV of signing a three-year lease for 100,000 sq. ft. of warehouse space is thus
NPV(Lease) = $38,364.The manager, however, still prefers to sign the three-year lease for
100,000 sq. ft. because this option has a higher expected profit.

Latihan
Sebuah perusahaan handphone melayani dua buah pasar yaitu pasar Asia dan pasar Amerika Utara.
Saat ini permintaaan dari pasar Asia adalah 2 juta handphone pertahun; sedangkan permintaan pasar
Amerika Utara adalah 4 juta handphone pertahun.. Selama dua tahuan kedepan akan terjadi peningkatan
permintaan Asia sebesar 50% dengan peluang sebesar 0,7 atau peningkatan permintaan Asia sebesar 20
% dengan peluang sebesar 0,3. Dalam periode yang sama, akan terjadi peningkatan permintaan dari
pasar Amerika Utara sebesar 10% dengan peluang sebesar 0,5 atau terjadi penurunan permintaaan dari
pasar Amerika Utara sebesar 10% dengan peluang sebesar 0,5.
Saat ini, Asia mempunyai fasilitas produksi dengan kapasitas 2,4 juta handphone pertahun dan Amerika
Utara mempunyai fasilitas produksi dengan kapasitas 4,2 juta handphone pertahun. Biaya variabel untuk
memproduksi handphone di Asia adalah $15 per unit ; sedangkan di Amerika Utara adalah $17 per unit.
Setiap handphone dapat dijual dengan harga $40 disetiap pasar. Anda diminta untuk:
Menghitung mana yang lebih baik: apakah menambah kapasitas sebesar 2 juta unit di Asia atau
menambah 1,5 juta unit di Amerika Utara. Penambahan pabrik di Asia akan membutuhkan biaya
investasi sebesar $18 juta; sedangkan penambahan pabrik di Amerika Utara membutuhkan biaya investsi
sebesar $ 15 juta. Discount factor 10%. Gunakan decision tree diagram untuk menyelesaikan
permasalahan ini.