Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
ELEMENT OF DECISION
PROBLEMS
Larkin Oil
Bill Mills shuffled his feet. The Spill Oil
Identify objectives
and alternatives
Decompose and
model the problem
Choose the best
alternative
A
Sensitivity Analysis
Is further analysis
needed?
N
Implement the
chosen Alternative
one begin?
A critical first step is that of identifying the
elements of the situation.
Uncertain events
The value of specific outcomes
Decision to make
Imagine a farmer whose trees are laden with
protective action
A decision: at least two alternatives
There may be a wide variety of alternatives
He may have several strategies for saving the crop
For example: wait and obtain more information But,
there may be not enough time to take action. The later
the decision, the worse outcome might be obtained.
Another possibility: Taking out insurance.
And also: Taking no action
Sequential decisions
In many cases, there simply is no single
Sequential decisions
Second
Decision
Third
Decision
Now
TIME LINE
Last
Decision
Uncertain events
In the previous discussion:
the decision problem can be complicated because
Example:
In a stock market, one investor will buy some
Resolved
Resolved
before second before third
decision
decision
Resolved after
last decision
Uncertain
Events
First
Decision
Second
Decision
Third
Decision
Now
TIME LINE
Last
Decision
Resolved
before last
decision
Resolved after
last decision
Uncertain
Events
Outcomes
First
Decision
Last
Decision
Second
Decision
Planning
Horizon
Now
TIME LINE
would be no problem.
In general, we talk about the present value of an amount
x that will be received at the end of n time periods
Location
Accident
Weather
Cause
Policy
Decision
Cost
Environmental
damage
Action
Management
Decision
Now
TIME LINE
Outcomes
Cost
Environmental
damage
PR damage
TI_Unand, 2012
Payoff ($ Billion)
Accept $2 Billion
2
5
10.3
Counter over
$5Billion
Final Court
Decision
Texaco Refuses
Counterover
Texaco
Counterover
$3 Billion
Refuse
10.3
Final Court
Decision
5
0
Accept $3 Billion
Payoff ($ Billion)
Accept $2 Billion
2
5
(0.2)
Counter over
$5Billion
Texaco Refuses
Counterover (0.50)
Texaco
Refuse
Counterover
$3 Billion (0.33)
10.3
5
(0.3)
(0.2)
10.3
(0.3)
Accept $3 Billion
Keep Ticket
Win 25
(0.20)
24
Lose
(0.80)
-1
Win 10
(0.45)
10
Lose
(0.55)
$4
Ticket
Keep
Ticket
$4.5
Texaco vs Pennzoil
EMV (Court Decision) =
[P(Award=10.3)x10.3] + [P(Award=5)x5]+ [P(Award=0)x0]
Texaco vs Pennzoil
Expected Value
Accept $2 Billion
2
(0.17)
Counter offer
$5Billion
Texaco Refuses
Counteroffer (0.50)
Texaco
Refuse
Counteroffer
$3 Billion (0.33)
4.56
4.56
Accept $3 Billion
Texaco vs Pennzoil
Expected Value
Accept $2 Billion
2
5
Counter offer
$5Billion
4.56
(0.50)
(0.33)
Accept $2 Billion
Counteroffer $5 Billion
Expected Value
2
4.63
Texaco vs Pennzoil
Payoff ($ Billion)
Accept $2 Billion
2
Counter over
$5Billion
4.63
Texaco Refuses
Counterover
(0.50)
Texaco
Refuse
Counterover
$3 Billion (0.33)
4.56
5
(0.2)
10.3
5
(0.3)
(0.2)
10.3
4.56
Final Court (0.5)
Decision
(0.3)
Accept $3 Billion
TI Unand, 2014
Pendahuluan
Mengapa pengambilan keputusan penting
Contoh persoalan keputusan
Tujuan Pembelajaran:
Mahasiswa mampu memodelkan struktur persoalan
keputusan (1 kriteria) agar dapat diambil keputusan
yang secara empiris lebih baik
Mahasiswa dapat menggunakan beberapa konsep
pengambilan keputusan multikriteria
Contoh Persoalan
Persoalan Texaco vs Penzoil
Awal Tahun 1984, Penzoil dan Getty Oil setuju untuk
merger. Tetapi sebelum dokumen formal
ditandatangani, Texaco menawarkan kepada Getty Oil
harga yang lebih baik. Pimpinan Getty setuju dengan
penawaran Texaco.
Penzoil yang merasa dirugikan, ingin menuntut Texaco
ke pengadilan.
Namun Texaco bersedia membayar ganti rugi $ 2 juta
Apa keputusan yang harus diambil oleh Penzoil? Terima
tawaran $ 2 juta tersebut atau tetap lanjutkan ke
pengadilan dengan keputusan final yang belum
diketahui?
Hasil ($ juta)
Terima $ 2 juta
5
Texaco Setuju $ 5 juta
5
Minta
$ 5 juta
10.3
Texaco Menolak
$ 5 juta
Putusan Final
Pengadilan
5
0
Texaco Menawar
$ 3 juta
Tolak
$ 3 juta
10.3
Putusan Final
Pengadilan
5
0
Terima $ 3 juta
Bahan Baku
Pasar
Tenaga Kerja
Infrastruktur
Lokasi 1
Lokasi 2
Lokasi 3
Lokasi 4
Langkah Pengambilan
Keputusan
1. Model struktur
persoalan
2. Model
ketidakpastian
3. Model preferensi
1. Identifikasi
Permasalahan
6. Implementasi
alternatif
terpilih
2. Identifikasi
obyektif dan
alternatif
5. Analisis
Sensitivitas
3. Dekomposisi
dan modelkan
persoalan
4. Pemilihan
Alternatif
Pembelajaran
MingMateri
gu
1 Pengantar
2
3
4
5
6
7
8
Pemodelan Keputusan:
Unsur-unsur dalam
Pengambilan Keputusan
Penstrukturan
Keputusan
Penetapan Pilihan
Analisis Sensitivitas
Presentasi I
Presentasi I (lanjutan)
UTS
MingMateri
gu
9 Kreativitas dalam
Pengambilan Keputusan
10 Perilaku Resiko dalam
Pengambilan Keputusan
11 Obyektif yang bertentangan:
Konsep Dasar
12 Model Multi Atribut
13 AHP
14 Presentasi II
15 Presentasi II (lanjutan)
16 UAS
Evaluasi:
1. Tugas
2. Aktivitas dalam pembelajaran
3. Presentasi
4. UTS
5. UAS
Referensi:
Clement, R.T. 1992. Making Hard Decision. PWSKent Publishing Company, Boston
Kesepakatan:
Kehadiran: 80%
Toleransi Keterlambatan: 15 menit (minggu 2 dan 3)
Pakaian-rambut:
No T-shirt
No trousers for women
No long hair for men
SENSITIVITY ANALYSIS
TI-Unand, 2014
Eagle Airlines
Identify
The
Decision to Make
Uncertain Event
The Outcomes
Influence Diagram
Operating
Cost
Hours
Flown
Insurance
Total
Cost
Capacity of
Scheduled
Flight
Revenue
Charter
Price
Price
Financial
Cost
Interest
Rate
Ratio
Charter/
Scheduled
Ticket
Price
PROFIT
Proportion
Financial
Purchase
Seneca?
Base Value
Lower Bound
Upper Bound
Hours Flown
800
500
1000
Charter Price/Hour
$325
$500
$550
Ticket Price/Hour
$100
$95
$108
Capacity on
Scheduled Flights
50%
40%
60%
Ratio of Charter to
Scheduled Flights
50%
45%
70%
Operating
Cost/Hour
$245
$230
260
Insurance
$20,000
$18,000
$25,000
Proportion
Financed
0.40
0.30
0.50
Interest Rate
11.5%
10.5%
13%
Purchase Price
$87,500
$85,000
$90,000
Tornado Diagram
Capacity on Scheduled Flights
Operating Cost/Hour
Hours Flown
Charter Price/Hour
Ratio of Charter to Scheduled Flights
Ticket Price/Hour
Insurance
Proportion Financed
Interest Rate
Purchase Price
-15000
-10000
-5000
5000
10000
15000
Expected Profit
20000
25000
30000
35000
0.55
Base
value
0.5
0.45
Profit <4200
0.4
230
235
240
245
250
255
260
Operating
Cost
SENSITIVITY TO
PROBABILITIES
Capacity of
Scheduled Flights
45%
(q)
$253
(p)
55%
(1-q)
Hours
Flown
650 (r)
-9,725
900 (1-r)
-4,225
650 (s)
6,525
900 (1-s)
650 (r)
Purchase
Piper
Seneca
$237
(1-p)
45%
(q)
55%
(1-q)
Do Not Purchase
Earn 8% of $52,500
Payoff ($)
18,275
675
900 (1-r)
10,175
650 (s)
16,925
900 (1-s)
32,675
4,200
Operating
Cost
Capacity of
Scheduled Flights
STRATEGY
REGIONS
45%
(q)
$253
(0.5)
55%
(1-q)
Purchase
Piper
Seneca
Hours
Flown
650 (r)
-9,725
900 (1-r)
-4,225
650 (0.8r)
6,525
900 (1-0.8r)
650 (r)
45%
(q)
$237
(0.5)
55%
(1-q)
Do Not Purchase
Earn 8% of $52,500
Payoff ($)
900 (1-r)
18,275
675
10,175
650 (0.8r)
16,925
900 (1-0.8r)
32,675
4,200
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Purchase Seneca
TI-Unand, 2014
Objective:
To show the roles of influence diagrams and
Decision trees
Elements
Decision
Representation
Chance event
Decision
nodes
Chance nodes
Value
Value nodes
If she invests,
she may be able to get in on the ground floor of a
Invest?
Value
Market
Activity
Outcome
Market Up
Market Down
Investment
choice
Payoff
Choices:
Choice:
Stock
market
outcome
Savings
Savings
Up
200
Down
200
Up
500
Down
-400
Stocks
Stocks
Payoff
Possible Reports:
Engineers
Report
Machine 3
OK?
Outcome
3 OK
I think, 3 is OK
3 Not OK
Managers
Decision
Choices:
Change Products
Replace 3
Payoff
Choice:
Outcome Payoff
Change
Products
3 OK
Replace 3
3 OK
3 Not OK
On schedule, costly
3 Not OK
Behind schedule
Possible Forecast:
Forecast
Hurricane
Path
Outcome
Hits Miami
Misses Miami
Payoff
Decision
Choices:
Choice:
Outcome
Evacuate
Evacuate
Hits Miami
Stay
Misses Miami
Stay
Payoff
Safety, high cost
Hits Miami
Misses Miami
Revenue
Introduce
Product?
Cost
Profit
Price
Introduce
Product?
Units
Sold
Fixed
Cost
Profit
Variable
Cost
Price
Introduce
Product?
Units
Sold
Fixed
Cost
Revenue
Cost
Profit
Variable
Cost
Bomb-detection
System Choice
Detection
effectiveness
Time to
implement
Passenger
Acceptance
Overall
satisfaction
Cost
Weather
Day 1
Weather
Day 2
Weather
Day n
Forecast
Day 1
Forecast
Day 2
Forecast
Day n
Protect
Day 1?
Protect
Day 2?
Protect
Day n?
Payoff
Day 1
Payoff
Day 2
Total
Payoff
Payoff
Day n
Usage
Decision
Economic
Value
Cancer
Cost
Net
Value
Usage
Decision
Economic
Value
Human
Exposure
Cancer
Risk
Carcinogenic
Potential
Cancer
Cost
Net
Value
Venture succeed
Invest
Venture fails
Do not Invest
Fund lost
US Representative
(Intermediate)
Win
US Senator
(Best)
Lose
Lawyer
(Worst)
Run for
Senate
Double-Risk Decision
Run for
Reelection
Win
Lose
Win
US Representative
Small-time lawyer
US Senator
Run for
Senate
Lose
Big-time lawyer
Known Amount
Highest
Reject
settlement
Lowest
Imperfect Information
Hurricane
Danger, Low cost
Hits Miami
Stay
Forecast: Will
Hit Miami
Evacuate
Hurricane
Hits Miami
Forecast: Will
Miss Miami
Stay
Hurricane
Safety, Low cost
Misses Miami
Evacuate
Job Offer 2
Madison, Wisconsin:
Low Salary
Chicago, Illinois:
High Salary
Sequential Decision
Forecast
Day 1
Decision
Day 1
Weather
Day 1
Forecast
Day 2
Decision
Day 2
Weather
Day 2
Total
Payoff