You are on page 1of 48

Master of Business Administration ( MBA ) Track: Supply Chain Management

Modeling & Decision Support System (ITS642)


Ahmad Magdy, MBA akabil@ltu.edu
1

Course Goals
I. Gain an understanding of the fundamental concepts of the managerial decision making process life cycle and problem solving strategies II. Learn how to gather, analyze, and represent data to describe managerial problems in an organizational environment III. Gain an understanding of the design and development strategies of DSS IV. Be able to elicit, identify, recognize, and capture user decision requirements for a DSS development V. Contribute as a group member in DSS development projects 2

Textbook
Decision Support and Business Intelligence Systems, 8/E Efraim Turban, et al. Copyright: 2007
Book Site: http://www.prenhall.com/turban

Decision Support Systems in the 21st Century 2nd Edition by George Marakas; Prentice-Hall, 2003.

Decision Making Programs, Kabeil, M., University of Sharjah, UAE, 2008.

Useful Links
http://www.dss-int.com http://businessintelligence.ittoolbox.com/

http://www.datawarehouse.com/
Textbook

site (http://www.prenhall.com/turban) Mitchell Face Recognition ANN UMBC Agent Site Navy Applied AI Lab http://www.uky.edu/BusinessEconomics/dssakba /instmat.htm Lots of DSS examples http://dssresources.com more DSS resources
6

Course Project
Each

group of 2-3 students will select a DSS application that incorporates the concepts discussed in the course. The outcomes of the application will be in the form of: - Report - Presentation

Course Project
The

course project will be a requirements analysis for a decision support system. Your group may select any decision support context you find most interesting. It could be for commercial use, military, or general governmental use. You will develop a requirements document that describes:

the decision process your DSS will support, a cost/benefits justification of the system, how it is projected to impact the organizational process it is designed to support, and a demonstration level/or working prototype of the system.

Grading System
Class

Work (Total 40 % of the final grade):

15 % Attendance and Class Participation.


10 % Group Project Report.

15 % Group Project Presentation.

Midterm Exams: 30 % of the final grade.

Final Exam: 30 % of the final grade.


9

QUESTIONS AND COMMENTS

10

Lecture 1 Introduction to DSS

11

The Secret to Happiness

12

DSS Overview
Who

? When ? Why ? What ? How ?

Who needs DSS ? When to use DSS ? Why we need DSS ? What is DSS ? How does DSS work ? -------------

13

What ?
is Decision What is Support What is Systems What is a Decision Support System
14

What

From Data to Decision

Source: Kabeil, Decision Making Programs, University of Sharjah, UAE, 2008. 15

The Decision Makers


Individuals Groups

16

Why are Decisions So Hard?


The four key areas that determine the relative difficulty of a decision are:
1.

Structure in general, the more structure, the less information required Cognitive limitations the human mind is limited to handling 5 to 9 distinct pieces of information [Experiment]
17

2.

Why are Decisions So Hard? (cont.)

3.

Uncertainty the amount is based on how complete and accurate the information is Alternatives and multiple objectives the selection of one alternative may impede the progress towards a different goal

4.

18

Simons Model of Problem Solving


Simon proposed a three-phase model of problem solving:
1.

2.

3.

Intelligence phase the decision maker looks for indications that a problem exists Design phase alternatives are formulated and analyzed Choice phase one of the alternatives is selected and implemented
19

The Decision-Making Process


Reality
Simplification
Assumptions

Intelligence Phase
Problem Statement

Validation of the model

Design Phase
Alternatives

Success Verification & testing

Choice Phase
Solution

Implementation
Failure

20

20

Decision-Making Phases

Intelligence Phase

Reality is examined Problem is identified, defined, and classified Problem ownership is established Making assumptions to simplify reality Write relationships among all variables Validate the model Set criteria for evaluation of alternatives Selection of solution to the model (not the real problem) Test and validate Solving the real problem

Design Phase

Choice Phase

Implementation

Failure leads to a return to an earlier phase

21

21

Rational Decision Making


Many

decision strategies (next slide) attempt to find optimal solutions. In many circles, this is considered to be rational behavior. It is not always possible to optimize. Some problems have only qualitative solutions. Others may be quantitative but have multiple objectives at odds with others. In such situations, rational behavior would be to choose a good solution.

22

Modeling and Analysis Strategies


Satisficing strategies: Optimizing strategies: Simulation Linear programming Forecasting Goal programming What if analysis Simple queuing models Environmental impact Investment models analysis Inventory models Transportation models

23

Bounded Rationality
Simon

argued that people dont always optimize because it is often impractical to consider all possible solutions to a problem. He notes that we often simplify reality by looking for a solution that is acceptable, a strategy he called satisficing. When people make rational decisions that are bounded by often uncontrollable constraints, he notes that they are operating inside bounded reality.

24

Optimizing Versus Satisficing

Optimizing strategies search here

Satisficing strategies search here


25

The Process of Choice


In

Simons model, the choice phase represents the climax of the decision process. It is important, however, to NOT focus all the energy here because it will not do justice to the other phases. The choice phase focuses mainly on decisions of the semistructured and unstructured types where there is uncertainty.
26

Cognitive Processes
Decision makers face a formidable task if they are to overcome all the factors contributing to cognitive limitations:
Humans can only retain a few bits of information in short-term memory. Decision makers display differing intelligence. Some decision makers tend to restrict their search. Decision makers that employ concrete thinking tend to be limited information processors.

27

Cognitive Processes (cont.)


Propensity for risk varies among decision makers. Decision makers level of aspiration is positively correlated with desire for information. In general, older decision makers appear to be more limited than younger ones.

28

Perception
This is a special type of cognitive limitation. Common perception blocks are:
Difficulty in isolating the problem. Delimiting the problem space too closely. Inability to see the problem from different perspectives. Stereotyping. Cognitive saturation or overload.

29

Stereotyping

30

Judgment
Although

numerous strategies exist for evaluation of solution alternatives, judgment appears to be the most favored. Compared to detailed analysis, judgment is faster, more convenient, and less stressful. When applied in isolation, however, judgment may be nothing but a guess. One reason why it may not be used exclusively is that it relies heavily on the decision makers recollection (memory), which may fail.
31

Biases and Heuristics In Decision Making


We

all have rules of thumb that we rely on in making decisions. Another term for such rules is heuristics. Heuristic search techniques follow a series of steps based on rules developed by experience. These searches can often provide solutions very close to those found by exhaustive search.
32

Definitions of Heuristic on the Web:


The word comes from the Greek word for discovery. A rule of thumb based on experience or common sense and incorporated in a knowledge base that is used to guide a problemsolving process A method employing experimentation, evaluation, common sense, trialand-error methods to learn, discover, understand, or solve problems.

Flexible guidelines or suggestions for solving a problem; not a step-bystep procedure (algorithm).
A process of solving problems by evaluating each step in the progress, searching for satisfactory solutions rather than optimal solutions

33

Advantages of Heuristics in Problem Solving


Simple

to understand Easy to implement. Requires less conception time. Requires less cognitive effort. Can produce multiple solutions.

34

Appropriate Uses of Heuristics in Problem Solving


Input

data are inexact or limited. High computation time for an optimal solution. Problems are solved frequently and repeatedly. Symbolic processing is involved. A reliable, exact method is not available. Optimization is not economically feasible.

35

Traveling Salesman Problem (Shortest Route) 9 Cities to Visit (just once)

Algorithm:
Check all possible combinations = 0.5(# of Cities-1)! = 0.5 (9 - 1)! = 20,160
36

Decision Making in Supply Chain Network Design

37

Solution Techniques
Mathematical

optimization techniques:

1. Heuristics: find good solutions, not necessarily optimal 2. Exact algorithms: find optimal solutions
Simulation

models: provide a mechanism to evaluate specified design alternatives created by the designer.

38

Example
Single Two

product

plants p1 and p2

Plant p2 has an annual capacity of 60,000 units.

The

two plants have the same production costs.

There There

are two warehouses w1 and w2 with identical warehouse handling costs.

are three markets areas c1,c2 and c3 with demands of 50,000, 100,000 and 50,000, respectively.

39

Unit Distribution Costs


Facility warehouse p1 p2 c1 c2 c3

w1
w2

0
5 P1

4
2

3
2

4
1

5
2

$0

W1

$3 $4 $2

D1 = 50,000

$5 $5 $4 $2

D2 = 100,000
$1

P2 Cap = 60,000

W2

$2

D3 = 50,000 40

Heuristic #1:Choose the Cheapest Warehouse to Source Demand

Heuristic #2: Choose the warehouse where the total delivery costs to and from the warehouse are the lowest [Consider inbound and outbound distribution costs]

41

The Optimization Model


The problem described earlier can be framed as the following linear programming problem. Let x(p1,w1), x(p1,w2), x(p2,w1) and x(p2,w2) be the flows from the plants to the warehouses. x(w1,c1), x(w1,c2), x(w1,c3) be the flows from the warehouse w1 to customer zones c1, c2 and c3. x(w2,c1), x(w2,c2), x(w2,c3) be the flows from warehouse w2 to customer zones c1, c2 and c3

42

Simulation Models
Useful

for a given design and a micro-level analysis. Examine:


Individual ordering pattern. Specific inventory policies. Inventory movements inside the warehouse.

Not

an optimization model Can only consider very few alternate models


43

Which One to Use?


Use

mathematical optimization for static analysis Use a 2-step approach when dynamics in system has to be analyzed:
Use an optimization model to generate a number of least-cost solutions at the macro level, taking into account the most important cost components. Use a simulation model to evaluate the solutions generated in the first phase.

44

Heuristic Bias
Sometimes the use of heuristics can hamper finding a solution. The four major categories of bias are: 1. Availability people tend to estimate probability based on past experience, which may not be representative. 2. Adjustment and anchoring people often pick a starting value and then adjust up and down from it. They tend to underestimate the need for adjustments.
45

Heuristic Bias (cont.)


Sometimes the use of heuristics can hamper finding a solution. The four major categories of bias are: 3. Representativeness people tend to misestimate probabilities of belonging to a group. 4. Motivational incentives often lead decision makers to estimate probabilities that do not reflect their true beliefs.
46

2-13: Effectiveness and Efficiency Effectiveness of DSS:


Easier access to information Faster problem recognition and identification Easier access to computing tools Greater ability to evaluate large choice sets

Efficiency from DSS:


Reduction in decision costs Reduction in decision time Better quality in feedback supplied

47

QUESTIONS AND COMMENTS

48

You might also like