REVIEW NOTES
Expected demand (mean or average value) = average simulated demand over a large #
of iterations
know which cells are the distribution (# demanded), parameter cell (# ordered), and
results cell
the disadvantage of simulation: each simulation
ASSIGNMENT #2: Decision Analysis
Due: Friday, November 7th, 2014 by 3:15 pm
in MGST 391 drop box (#26)
Note: If there is a discrepancy in values between the Word file and the Excel templates, use
the Excel template values.
1.
Plaid Renovations Company [30
DECISION ANALYSIS
All decision making problems have 3 elements:
1. alternatives for decision maker
2. possible outcomes and probabilities of these outcomes
3. a value model that shows results for various combinations of these outcomes
* Once these element
QUEUING TEMPLATES
QUEUING TEMPLATES
2005 by David W. Ashley
2005 by David W. Ashley
Ver. 9.3.01
Ver. 9.3.01
This workbook computes queuing results for the following models:
M/M/s
M / M / s with finite queue length
M / M / s with finite arrival populatio
A survey of ideas, trends, people, and practices on the business horizon
g r i st
Suppose that an executive wants to make
a prediction that bears on his company.
Will a particular product sell? How will
a particular job applicant perform?
When will a new
+
Module 1 Lab
Introduction to Excel
Functions & Influence
+
2
Name Card
FirstName
LastInitial
Fold
in half
FirstName
LastInitial
MGST 217
Haskayne School of
+
3
Excel Skills and Functions Module 1
Titlebar
Tabs
Ribbons
Worksheet Tabs
Name Box
Formula Bar
+
Module 1:
Solving Business
How to approach
Problems
business problem
+
2
Why are you here?
n
Learn about structured ways to look at business problems
n
Learn about ways to present business solutions
MGST 217
Haskayne School of
+
3
Case framework for Eth
+
MGST 217 Day1
What is MGST 217 all
About?
+
2
Why are we here?
n
To get to know each other a bit
n
Set our classroom expectations
n
Learn about MGST 217 and why you are required to take it
MGST 217
Haskayne School of
+
3
About Us
n
Meet a neighbour or t
+
Module 2:
How much profit can I
Modelling your
make?
accounting needs
+
2
Administrivia
n
THRIVE Program Video
n
Career Carnival
n
TopHat Questions
MGST 217
Haskayne School of
+
Learning Outcomes For Module 2
n
Create an INFLUENCE DIAGRAM with the prope
Integer Linear Programming

Use a binary constraint in solver when the changing cell is talking on a yes/no
function

Use integer constraint when partial amount cant be used (eg. partial ads cant be
purchased)

Cant get a sensitivity report with intege
QUEUING
o
Waiting from: not enough servers (s), slow servers (mu), too many customers
(lambda), customers before you and cant cut line (FIFO)
o
What people do in line ups: wait, join, not join, join then leave line up, cut in,
jockey (change lines), or me
M/M/1
o
Poisson arrival process, exponential service times, and 1 server
o
Good model where customers arrive without appointments (ex.ATM)
o
Customers join single line up for one server
o
Interarrival times and service times are exponential
o
o
FIFO servi
FORECASTING
o
Forecasting is always wrong (just depends how much)
o
Should include a measure of error (how far off) to determine which forecasting method is
best
o
Short range forecasts are more accurate than long range forecasts
o
Using past as indicatio
SIMULATION
2 major categories of simulation:
Monte Carlo= repeated samplings to characterize distribution of output; usually done on
computer
Systems=models the dynamics of interacting elements; often associated with physical
simulations
Simulation= devel
SIMPLE MOVING AVERAGE FORECASTS
A) Two Period Moving Average:
Month
Jan
Feb
Mar
Apr
May
Jun
Actual Demand (Y(t)
1500
1480
1520
1600
1650
1750
Forecast (F(t)
1490
1500
1560
1625
Error (e(t) = Y(t)F(t)
30
100
90
125
B) Three Period Moving Average:
Month
Ja
Nature of Decision Making Supplementary Readings
Day 1: Individual Biases in Decision Making
Tversky, A. and Kahneman, D. (Sept. 2, 1974). Judgment under Uncertainty: Heuristics and
Biases. Science, New Series, vol. 185, pp. 11241131. Published by: Ameri
SIMPLE MOVING AVERAGE FORECASTS
A) Two Period Moving Average:
Month
Jan
Feb
Mar
Apr
May
Jun
Actual Demand (Y(t)
1500
1480
1520
1600
1650
1750
Forecast (F(t)
Error (e(t) = Y(t)F(t)
B) Three Period Moving Average:
Month
Jan
Feb
Mar
Apr
May
Jun
Actual Deman
TERRAPE ENERGY INSURANCE RISK
Data:
Alternative
Policy
1
2
3
4
A
B
C
None
Deductible
$10,000
$25,000
$50,000
$0
Annual
Premium
$75,000
$50,000
$30,000
$0
Frequency of Occurrence
10.0%
Damage Extent: Mean $450,000
Damage Extent: Std. Dev. $50,000
Analysis:
Terrape Energy DRILLING CHOICES
Well ID#
Est. Recoverable barrels (000,000's)
Budgeted Cost ($ millions)
Proceed with well site? (0No, 1Yes)
AB1
5
$30
0
Total Recoverable Barrels (000,000's)
39
CONSTRAINTS
Total # of wells drilled
No more than 2 in Sask
TERRAPE ENERGY INSURANCE RISK
Data:
Alternative
Policy
1
2
3
4
A
B
C
None
Deductible
$10,000
$25,000
$50,000
$0
Annual
Premium
$75,000
$50,000
$30,000
$0
Frequency of Occurrence
10.0%
Damage Extent: Mean $450,000
Damage Extent: Std. Dev. $50,000
Simulatio
Study Guide for Nature of Decision Making (30% of Midterm)
1st class = Biases in Decision Making
Heuristics
o
o
Representativeness heuristic
o
Availability heuristic
Anchoring and Adjustment
Cognitive biases
o
Bias due to the Effectiveness of a Search Set
TERRAPE ENERGY INSURANCE RISK
Data:
Alternative
Policy
1
2
3
4
A
B
C
None
Deductible
$10,000
$25,000
$50,000
$0
Annual
Premium
$75,000
$50,000
$30,000
$0
Frequency of Occurrence
10.0%
Damage Extent: Mean $450,000
Damage Extent: Std. Dev. $50,000
Analysis:
CYCLICAL FORECASTING
2008 Total Forecast:
186,000 2008 Monthly Forecast (= D3/12):
Historical Demand
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
TOTAL
2006
12,124
12,626
14,082
14,668
16,527
17,223
18,534
17,716
16,117
14,383
13,572
12,178
179,750
Ave
CYCLICAL FORECASTING
2008 Total Forecast:
186,000 2008 Monthly Forecast (= D3/12):
Historical Demand
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
TOTAL
2006
12,124
12,626
14,082
14,668
16,527
17,223
18,534
17,716
16,117
14,383
13,572
12,178
179,750
Ave