Post on 18-Jan-2021
Introduction to Decision Making
Dr. Ioannis N. Lagoudis ilagoudis@misi.edu.my
lagoudis@mit.edu
Problems • Personal: – Job – Car purchase
• Community/local/national: – Drugs – Road/Airport/Hospital construction
Decision Making Process
Understanding the Problem
Formula4ng the Problem
Uncertainty Analysis U4lity Analysis
Op4miza4on
Understanding the Problem • Defining the problem • Identifying the stakeholders in the decision
making process • Knowing the available alternatives
Formulating the Problem • Identify the events • Document the information that is available: – From the beginning – At a later stage
• Visualize events and information based on time.
Uncertainty Analysis • Mainly with the use of probabilities • Probability estimation is critical • Probabilities can be derived based on: – Historical data – Expert opinion – Subjective estimation of stakeholders involved in
the decision making process
Utility / Value • Identify values to alternatives: – Objective – Subjective
Optimization • Estimate the different alternatives • Compare among the different alternatives • Use sensitivity analysis to evaluate the
robustness of the alternatives • Choose the “best” alternative
Types of Problems • There are two broad categories: – Simple – Complex
Frequent Problems during Decision Making
• Identify the variables that are of importance • Filtering the right variables • Variable evaluation • Number of involved stakeholders
Decision Making Tools • Statistics Regression Analysis • Structural Equation Modeling (S.E.M.) • Analytic Hierarchy Process (A.H.P.) • Multi-attribute Utility Theory (M.A.U.T.) • Decision Tree Model • Simulation • Linear Optimization
Statistics • Mean • Median • Quartiles • Probabilities • Histogram
• Variance • Standard Deviation • t distribution • Hypothesis testing • Sample
Regression
• Simple
– y=β0+β1χ1
• Multiple – y=β0+β1χ1+β2χ2+…+βnχn+ε
110 xy ββ +=
y
x
β1
β0
Structural Equation Modeling
• Allows the comparison of independent variables
LatentVariable“A”
LatentVariable“A”
LatentVariable“B”
LatentVariable“B”
I1I1 I2I2 I3I3 I4I4 I5I5 I6I6
δ δ
λ λ
θ
LatentVariable“A”
LatentVariable“A”
LatentVariable“B”
LatentVariable“B”
I1I1 I2I2 I3I3 I4I4 I5I5 I6I6
δ δ
λ λ
θ
SecondOrderFactor“AB”
SecondOrderFactor“AB”
Γ
BasicBasic Second OrderSecond Order
LatentVariable“A”
LatentVariable“A”
LatentVariable“B”
LatentVariable“B”
I1I1 I2I2 I3I3 I4I4 I5I5 I6I6
δ δ
λ λ
θ
LatentVariable“A”
LatentVariable“A”
LatentVariable“B”
LatentVariable“B”
I1I1 I2I2 I3I3 I4I4 I5I5 I6I6
δ δ
λ λ
θ
SecondOrderFactor“AB”
SecondOrderFactor“AB”
Γ
BasicBasic Second OrderSecond Order
Analytic Hierarchy Process
GOALGOAL
Criteria (A)Criteria (A) Criteria (B)Criteria (B) Criteria (C)Criteria (C) Criteria (D)Criteria (D)
A1A1 AnAnA2A2 B1B1 BnBnB2B2 C1C1 CnCnC2C2 D1D1 DnDnD2D2
Level 1
Level 2
Level 3
Level 4 Alternative 1Alternative 1 Alternative 2Alternative 2 Alternative 3Alternative 3
Attributes
GOALGOAL
Criteria (A)Criteria (A) Criteria (B)Criteria (B) Criteria (C)Criteria (C) Criteria (D)Criteria (D)
A1A1 AnAnA2A2 B1B1 BnBnB2B2 C1C1 CnCnC2C2 D1D1 DnDnD2D2
Level 1
Level 2
Level 3
Level 4 Alternative 1Alternative 1 Alternative 2Alternative 2 Alternative 3Alternative 3
Attributes
Multi-Attribute Utility Theory
GOAL Criteria (a)
Criteria (c)
Criteria (b)
a1
a2
an
b1
b2
bn
c1
c2
cn
Decision Tree Models
Α
C
Β
Value in €
Value in €
Value in €
Value in €
Op4on Β
p1
p2
p3
p4
Simulation • Use of software • Allows for detailed formulation on problems • Enables scenario analysis via the creation of
random numbers • SIML8, VENSIM, ARENA, MICROSOFT
EXCEL
Linear Optimization • Optimization tool • Uses mathematical
modeling
• Max 130W + 100P – 1.5W + 1.0P ≤ 27 – 1.0W + 1.0P ≤ 21 – 0.3W + 0.5P ≤ 9 – W ≤ 15 – P ≤ 16 – W,P ≥ 0
Thank you!