Probability and Decision Trees

1
Probability theory and Probability theory and

Decision trees Decision trees
Probability Probability
Thequantificationof uncertainty The quantification of uncertainty
Uncertainty may also be expressed as
likelihood,chance or risk.
Probability is expressed by p and can only
takes values ranging from 0,ie impossibility, g g , p y,
to 1, ie certainty.
2
Objective and subjective Objective and subjective
probability probability
Objective probability
wheretheprobability
Subjective probability
where a probability value
where the probability
of an event is based on
past data and the
circumstances are
repeatable by test.
Eg. The probability of
tossingacoinanda
p y
cannot be tested by repeated
trials.
The quantification is based
on personal judgement,
experience and expertise.
Eg. a QS considers that
there is 75% chance of
tossing a coin and a
head showing is 50%
or 1/2 or 0.5.
winning a tender for which
his firm has just entered
Basic rules of probability Basic rules of probability
Multiplicationrule
Eg. What is the
Multiplication rule
(AND)
This rule is used when
there is a string of
independent events for
which each individual
probabilit iskno n
g
probability of throwing
a 3 AND a 6 with two
throws of a die?
p(throwing a 3) =1/6
probability is known
and it is required to
know the overall
probability
p(throwing a 3 AND a 6)
=1/6 x 1/6
=1/36
3
Additionrule(OR)
Eg. What is the
Addition rule (OR)
This rule is used to
calculate the
probability of two or
more mutually
exclusive events. In
suchcircumstancesthe
g
probability of throwing
a 3 OR a 6 with a throw
of a die?
such circumstances the
probabilities of the
separate events must be
added.
p(throwing a 3 OR a 6)
=1/6 +1/6
=1/3
Conditional probability
Thisisaprobabilityassociatedwithcombinationsof This is a probability associated with combinations of
events but given that some prior result has already been
achieved with one of them.
When the events are independent of one another, then
the conditional probability is the same as the probability
of the remaining event
usuallyexpressedintheform: p(x|y) whichmeans usually expressed in the form: p (x|y) which means -
probability of x given that y has occurred
4
Eg. The probability of throwing a total of 10 with 2
dice, before the events, is 1/12; , , ;
ie. p(5 and 5) =1/6 x 1/6 =1/36
p(6 and 4) =1/6 x 1/6 =1/36
p(4 and 6) =1/6 x 1/6 =1/36
=3/36
=1/12
but if one of the die has been thrown and shows a 4
then the conditional probability is the probability of
throwing a 6 with the other die, which the probability
is 1/6
Example Example
Formpast experienceit isknownthat amachineis Form past experience it is known that a machine is
set up correctly on 90% of the occasions. If the
machine is set up correctly then 95% of good
products are expected but if the machine is not set
up correctly then the probability of a good product
is only 30%. On a particular day the machine is set y p y
up and the first product was found to be good.
What is the probability that the machine is set up
correctly?
5
Solution Solution
Probability tree
CSGP = 0.9 x 0.95 =0.855
CSBP = 0.9 x 0.05 =0.045
ISGP = 0.1 x 0.3 =0.03
CS =0.9
IS=01
GP =0.3
GP =0.95
BP=0.05
ISBP = 0.1 x 0.7 =0.07
IS =0.1
BP=0.7
Solution Solution
Probability of getting a good product
=CSGP +ISGP =0.855 +0.03 =0.885
Probability that machine is correctly set up
after getting good product is
CSGP = 0.855 = 0.966
CSGP +ISGP 0.885
6
Other example Other example
Thereare100final year studentsin There are 100 final year students in
Department of Quantity Surveying. 36 are
male and are studying qs, 9 are male and not
studying qs, 42 are female and studying qs, 13
are female not studying qs.
Th t ff ti t h dl thi d t The most effective way to handle this data
is is to draw up a table.
Gender and course table Gender and course table
QS
Not
QS
Total
MALE
M
FEMALE
QS
A
QS
A
Total
36 9 45
55
FEMALE
F
TOTAL
42 13
55
78 22 100
7
Expected value Expected value
Whereoptionshavevalues(somuchprofit Where options have values (so much profit,
contribution etc) as well as probabilities, the
concept of expected value is often used.
The expected value of an event is its
probability times the outcome or value of
the event over a series of trials.
Two projects are being considered and Two projects are being considered and
it is required to calculate the expected value it is required to calculate the expected value
of each project. of each project.
Project A Project B
RM p EV RM p EV
Optimistic 6000 x .2 =1200 6500 x .1 =650
Most likely 3500 x .5 =1750 4000 x .6 =2400
Pessimistic 2500 x .3 =750 1000 x .3 = 300
PROJ ECT EV RM 3700 RM 3350
8
Exercise Exercise
A firmmakingsanitaryfittingshavebeen A firm making sanitary fittings have been
considering the likely demand for fittings over
the next 6 years and think that the demand
pattern will be as follows:
High demand for 6 years p =0.5
Lowdemandfor 6years p=0.3 Low demand for 6 years p 0.3
High demand for 3 years followed
by Low demand for 3 years p =0.2
Exercise Exercise
Enlargement of capacity is required and the
followingoptionsareavailable: following options are available:
Option A Install fully-auto facility
immediately at a cost of RM5.4 m
Option B Install semi-auto facility
immediately at a cost of RM4m
Option C Install semi-auto facility
i di l i B d d immediately as in B and upgrade to
fully automatic at an additional cost
of RM2m in 3 years time providing
demand has been high for the 3 years
9
Exercise Exercise
Thereturnexpectedfor thevariousdemand The return expected for the various demand
and capacity options are estimated to be:
HIGH DEMAND LOW DEMAND
OPTION A RM1.6m RM0.6m
B RM0.9M p.a RM0.8m
then RM0.5 p.a
for 3 years
C RM 0.9m p.a RM0.8m p.a for 3
then RM1.1m years then RM0.3m
for 3 years p.a for 3 years
b
a
High Demand
1.6m p.a for 3 years
Low demand
06mpafor 3years
High Demand
1.6m p.a for
3 years
Years 1 - 3 Years 4 - 6
1
2
3
D1
b
D2
c
OPTION A
Fully auto
(cost 5.4m)
OPTION B
High Demand
16mpafor
OPTION C
Upgrade
(cost 2m)
Low demand
High Demand
Low demand
03mpafor 3years
4
1
2
3
e
D2
d
OPTION B
semi-auto
(cost 4m)
1.6m p.a for
3 years High Demand
Low demand
0.8m p.a for 3 years Low demand
1
2
4
10
Defining the required Defining the required
probability probability
) | ( 13 46 HD HD P
1
) (
) | (
) | (
13
13 46
13 46
HD P
HD LD P
HD HD P
2
3
) ( 13 LD P 4
Calculating the required Calculating the required
probs. probs.
Years 4 - 6
HIGH LOW HIGH
DEMAND
a
.5
LOW
DEMAND
b
.2
D
H
I
G
H
D
E
M
A
N
D
a
r
s

1

-
3
c
0
d
.3
H
I
G
H
D
E
M
A
N
D
Y
e
a
11
Calculating the required Calculating the required
probs. probs.
5 0 ) & ( HD HD P
0 ) & (
2 . 0 ) & (
5 . 0 ) & (
46 13
46 13
46 13
HD LD P
LD HD P
HD HD P
3 . 0 ) & (
) (
46 13 LD LD P
b
a
High Demand
Low demand
06mpafor 3years
High Demand
1.6m p.a for
3 years
Years 1 - 3 Years 4 - 6
1
2
3
0.7
0.71
0.29
D1
b
D2
c
OPTION A
Fully auto
(cost 5.4m)
OPTION B
High Demand
16mpafor
OPTION C
Upgrade
(cost 2m)
Low demand
High Demand
Low demand
03mpafor 3years
4
1
2
3
0.3
0.71
0.29
e
D2
d
OPTION B
semi-auto
(cost 4m)
1.6m p.a for
3 years High Demand
Low demand
0.8m p.a for 3 years Low demand
1
2
4
0.7
0.3
0.29
0.71
12
Analysis of decision tree Analysis of decision tree
The analysis is carried out out by rolling back from
node to node from RIGHT to LEFT, hence the term
BACKWARD PASS
EMV Node a=(0.71x1.6x3) +(0.29x0.6x3)
=3.408 +0.522
=RM3.93m
EMV Node b=(0.7((1.6x3)+3.93) +(0.3x0.6x6)
6111+108 =6.111 +1.08
=RM7.191m
Net EMV of OPTION A=7.191 - 5.4
=RM 1.791m
EMV Nodec=(071x11x3) +(029x03x3) EMV Node c=(0.71x1.1x3) +(0.29x0.3x3)
=RM2.604m
Net EMV =2.604 - 2 =RM 0.604m
EMV Node d=(0.71x0.5x3) +(0.29x0.8x3)
=RM1.761m
EMV Node D2:
EMV Node c EMV Node d
Decision at D2: Not upgrade from OPTION B to
OPTION C
13
EMV Nodee=(0.7((0.9x3) +1.761) EMV Node e (0.7((0.9x3) 1.761)
+(0.3x0.8x6)
=3.1227 +1.44
=RM 4.5627m
Net EMV of OPTION B =4.5627 - 4
=RM0.5627m
EMV Node D1:
EMV Node b = RM1.791m OPTION A
EMV Node e = RM0.5627m OPTION B
Decision at D1: To opt for OPTION A that is to install
the fully automatic machinery at the outset.

Probability and Decision Trees

Uploaded by

Copyright:

Available Formats

Probability and Decision Trees

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Probability and Decision Trees

Uploaded by

Copyright:

Available Formats

1

Probability theory and Probability theory and

You might also like