Welcome to Decision Support Systems

Saint Ambrose University
College of Business
MBA 782M Decision Support Systems
Fall 2005
(Date revised: October
3, 2005)

Class Meets:

Instructor:

Mondays 6:15 – 9:15 PM, McMullin 101

Bob Grenier, PhD, CCP, CDP
(H) 309.736.7335   (Augie) 309.794.7399
Dr.Grenier@mchsi.com
http://dr.grenier.home.mchsi.com

St. Ambrose University supports students' success by providing a broad range of reasonable accommodations for qualified students with disabilities. Services do not lower standards or alter degree requirements but instead give students a better chance to demonstrate their academic abilities. If you need an accommodation due to a disability, please contact the Office of Services for Students with Disabilities, at 319/333-6275 VC/TTY as soon as possible.

Guide to the syllabus

This syllabus is subject to change -- check it frequently! The current version is available on my web site at the following URL: http://dr.grenier.home.mchsi.com/SAU/syl782-2005.pdf.

Objectives

Text

Software

Learning Methods

Grading

Administrative

Assignments

Schedule

Bibliography

Prerequisites:

None

Course Objectives:

Decision support systems (DSS) are information systems that provide information to help knowledge workers make unstructured or semistructured decisions, i.e. decisions that require some human judgement.

Upon the successful completion of this course, a student should:

·        Demonstrate an understanding of DSS.

·        Understand related topics such as Representational Models, Optimization, Artificial Intelligence (AI), Expert Systems (ES), Neural Networks, and Fuzzy Logic.

·        Be able to differentiate between data based and knowledge based systems.

·        Be able to identify tools and techniques appropriate for problem solution.

Text:





Efrem G. Mallach

Decision Support and Data Warehouse Systems, 2000

Irwin McGraw-Hill, ISBN 0-07-289981-6

Supplementary Materials:

The Computer Information Center (http://www.compinfo-center.com) is a very useful portal is a variety of resources about information technology.

There are numerous online magazines devoted to computer technology and related issues. An extensive set of hypertext links is available at http://www.compinfo-center.com/itmags.htm.

The Computer Information Center  contains Knowledge Bases, Newsgroups and FAQs, Magazines and Ezines, White Papers, Organizations and User Groups, News, Events, Related Topics, etc. for concepts sometimes used with DSS such as:

Artificial Intelligence

Expert Systems

Intelligent Agents

Pattern Recognition -

Artificial Life

Fuzzy Logic

Knowledge Management

Robotics

Bayesian Belief Networks

Genetic Algorithms

Knowledge Representation

Supercomputers

Computer Vision

Genetic Programming

Machine Learning

 

Cybernetics

Industrial Control Systems

Neural Networks

 

The Web contains a wealth of information about subjects discussed in this course. However, the sheer volume can be overwhelming. Finding pertinent information is sometimes like finding the proverbial needle in the haystack. Before using information found on the WWW, I recommend you visit some of the sites listed on my web page Citing and Evaluating Resources found on the Internet   http://www.augustana.edu/users/bagrenier/resourcesusing.html

Software:

We will be using the following software:

Internet email will be used extensively for communication and collaboration. Be sure to check your email regularly each week. Students may submit questions and suggestions via email. I will email answers and responses individually or to the class as appropriate. I will also accept assignments as email attachments of Word, Excel, or Powerpoint files. If you do not have an ISP (Internet Service Provider), the University provides all students with accounts. Companies such as Juno offer free email service.

The Web will be used to access the current course syllabus, other pertinent materials, and also for research.

MS Excel(or equivalent) for constructing a DSS.

EXSYS Professional (student edition) for construction of an expert system. I have diskettes containing version 3.02, which runs in a DOS Window and is limited to 25 rules. You can also download a zipped executable file exsys.exe. To learn how to use the program run a:\demo. To construct an expert system, run a:\prodemo.

Software of your choice to draw class diagrams and other models (ERDs, etc.). If you do not have access to such software, I would suggest downloading Smart Draw from the Internet http://www.smartdraw.com

Learning Methods:

I offer the following suggestions for attaining the maximum benefit from the readings. Each chapter in the text begins with an Outline, Introduction, and Objectives and concludes with a Summary and a list of Key Terms. Read the objectives, introduction, and summary first, and then read the chapter. This practice will make the chapter more understandable on the first reading. When you are done with the chapter, review the key terms and concepts list. If any item is not fully understood look it up in the glossary at the end of the text or in the chapter as indicated.

Grading:

Assignment

Weight

Class/group participation

10%

Quizzes (best 5 out of 8)

10%

Case Study Analysis

10%

DSS – model using NPV

20%

DSS – aggregate planning using goal seeking

10%

Research Paper & Presentation

20%

Expert System

20%

Class/group participation
Participation in class discussion is of paramount importance! Be prepared by completing the assigned readings and assignments.

Quizzes
Quizzes will be given at the beginning of the class period. Generally the quiz will consist of 10 multiple choice questions about the reading assignment. The quiz will be corrected in class and the answers negotiated.

Administrative Considerations:

Assignments turned in, or received by email, after the beginning of class on the day in which they are due are subject to a 10% penalty. Since assignments can be submitted by email, absence from class is not a good reason for late assignments. Assignments will not be accepted if they are more than one week late. Quizzes must be taken when scheduled.

Group vs. Independent Work:

Most assignments are to be done individually, the exceptions being DSS and ES development projects. Although giving and receiving some help from other students is OK, avoid go so far as to essentially do someone's assignment for them.

Assignments:

Be sure to bring a copy of the assignment to class even if you have submitted a copy via email since it will be discussed during class.

Explanations of the assignments can be found at the end of this syllabus and can be accessed by following the hyperlinks below.

Schedule:

Meeting

Topic

Assignments Due

Reading

1
8/29

Introduction to DSS
Human Decision Making Processes

 

Ch.1 & 2 
DSS Intro
Decision Making

9/5

NO CLASS – Labor Day

 

 

2
9/12

Systems, Information Quality, & Models
Types of DSS

Quiz

Ch.1, 2, 3 & 4
pp. 1-160
Systems, Info....
DSS Types

3
9/19

DSS Architecture, Hardware, & Operating System Platforms

Quiz

Ch. 5
pp. 163-196
Architecture
Client/Server

4
9/26

DSS Software Tools

Quiz
Case 1

Ch. 6
pp. 197-256
Tools
Forecasting

5
10/3

Building and Implementing DSS

Quiz

Ch. 7
pp. 258-296
Development

6
10/10

Models in DSS

Quiz
Case 2

Ch. 8
pp. 297-344
Models

10/17

NO CLASS

 

 

7
10/24

Mathematical Models and Optimization

Quiz
DSS NPV

Ch. 9
pp. 346-383
Math Models

8
10/31

Group DSS

Quiz

Ch. 10
pp. 384-423
GDSS

9
11/7

Expert Systems

Quiz
DSS Capacity

Ch. 11
pp. 424-459

10
11/14

Data Warehousing

 

Ch, 12-14
pp. 465-556

11
11/21

NO CLASS – Thanksgiving

 

 

11/28

Research paper presentations

 

 

12
12/5

Expert System presentations

 

 

Bibliography

GENERAL

O'Keefe Library, St. Ambrose University (http://library.sau.edu/) - there are excellent resources available here including a number of indices and databases such as: Business ASAP, FirstSearch, Newsbank, and much more. Some resources are only available to Saint Ambrose students, i.e. you will need an id and password.

The Computer Information Center (http://www.compinfo-center.com).

PCAI – Where Intelligent Technology Meets the Real World - http://www.pcai.com. This site provides information, examples of successful implementations, tutorials, new product announcements, buying guides and more about intelligent applications, such as, Expert Systems, Fuzzy Logic, Neural Networks, Genetic Algorithms, etc.

ARTIFICIAL INTELLIGENCE (AI)

American Association for AI -- http://www.aaai.org/

Generation 5’s goal is to be the most comprehensive Artificial Intelligence site on the Web – http://generation5.org.

Association for Computing Machinery - Special Interest Group on AI (SIGART) AI Resources -- http://sigart.acm.org/ai/

Carnegie Mellon AI Repository --
http://www.cs.cmu.edu/afs/cs.cmu.edu/project/ai-repository/ai/0.html

EXPERT SYSTEMS (ES)

The Heynneman AI web site (http://www.heynneman.com) contains forty case histories of ES developed using EXSYS.

Frequently Asked Questions about ES -- including pointers to free/cheap ES shells
http://www.cs.cmu.edu/afs/cs.cmu.edu/project/airepository/ai/html/faqs/ai/expert/part1/faq.html

ES Books with disks -- http://www.geocities.com/SiliconValley/Lakes/6007/Expert_Books.html

Information and news on the CLIPS rule-based language developed by NASA's Johnson Space Center http://www.siliconvalleyone.com/clips.htm

CLIPS -- http://www.ghg.net/clips/CLIPS.html

Java ES Shell (like CLIPS) -- http://herzberg.ca.sandia.gov/jess/

Exsys CORVID – company that sells a descendant of the original Exsys expert systems development tool. The site has several demonstrations of actual applications, e.g. Cessna Diagnostic and Repair System, Immigration and Naturalization Service, Environmental Compliance Appraisal, Weather Instrument Product Selector, State of Wisconsin Public School Disciplinary Action Advisor, Camcorder Selector, and several others. Visit http://www.exsys.com/demomain.html

FUZZY LOGIC -- Fuzzy logic is precise reasoning about imprecise concepts.

Complete Repository for Fuzzy Logic Applications, including Application Notes, Simulation Software, and Teaching Materials -- http://www.fuzzytech.com/index.php

Fuzzy Logic Tutorial -- http://www.fuzzy-logic.com/index.htm

Case Study Analysis

The Appendix of the text contains nine case studies. Choose two cases and perform a case analysis for each one.

Your analysis should be presented in the following format:

 

The analysis should be prepared according to the following guidelines:

DSS development:

You may do this project by yourself or with a teammate. Notify me by the end of the second class meeting as to your choice.

You are planning to replace a legacy inventory management system with a new client-server system. The cash flows associated with a client-server system development project are as follows:

There is an initial cost of $120,000. In the subsequent 5 years the payoffs are all projected to be positive and are estimated to be $40,000, $60,000, $60,000, $50,000, and $40,000.

Using MS Excel, or an appropriate equivalent, develop a DSS to answer the following questions.

(a) Assuming that the interest rate is 10%, what is the Net Present Value (NPV) of this project? [4 points]

(b) Over what range of interest rates is the NPV of this project positive? [4 points]

(c) Instead of having to pay $120,000 up front, you are given the option of paying only $60,000 initially and an additional $70,000 the following year. Assuming an interest rate of 10%, would you accept this offer? Why? [6 points]

(d) The NPV (at 12% interest rate) of continuing with the legacy inventory management system is estimated to be $50,000. If the client-server system is successfully implemented, the net present value of the new system at the same interest rate is approximately $60,000. However, there is a 30% chance that the project will run into cost over-runs. Under this unfortunate scenario, the NPV of the client-server system is estimated to be only $10,000. Based on this information would a risk-neutral decision-maker replace the legacy system? Justify your answer. [6 points]

Deliverables include:

DSS – aggregate planning for Bradford Manufacturing

The case is described in the following html document: http://dr.grenier.home.mchsi.com/SAU/Bradford.html

The model is http://dr.grenier.home.mchsi.com/SAU/Bradford_Manufacturing.xls

Research Paper and Presentation:

Prepare a research report or product evaluation on one of the subject areas listed below according to the guidelines outlined for the Case Study Analysis. The report should be at least 5 pages in length. The report should address the following as appropriate

Each student will report on a different subject. During the second class meeting, a lottery will be held to determine the priority for choosing topics.

SUBJECT AREA

RESEARCHER

Artificial Intelligence

 

CRM

 

Data mining

 

Expert Systems

 

Fuzzy Logic

 

Genetic Algorithms

 

GDSS (Decision Support Systems)

 

Intelligent Agents

 

Knowledge Management

 

Neural Networks

 

OLAP (On-line Analytical Processing)

 

ES DESIGN AND CONSTRUCTION:

You may do this project by yourself or with a teammate. Notify me by the end of the second class meeting as to your choice.

Design and construct an expert system to address a problem of your choice. Before proceeding, please discuss the problem with me. You should use EXSYS or an approved equivalent. The deliverables for this assignment should be prepared below according to the guidelines outlined for the Case Study Analysis and include:

The resulting ES will be presented to the class during the last meeting.

Possible ES Applications

Ideally you should pick an application you are familiar with and which might be valuable in your business or personal life. Knowing an expert in the application is very helpful.

The projects described below are merely suggestions or starting points intended for those who need an idea. Feel free to make modifications you deem necessary.

Fitness evaluator

There are numerous resources (web sites, magazine and newspaper articles, etc.) that present questions to evaluate personal fitness. Questions may deal with life habits as well as personal information such as age, height, weight, blood pressure, etc.. Some resources include formulas or tables for calculating fitness indicators such as maximum pulse rate or BMI (Body Mass Index). Often recommendations are made about diet or exercise. The questions and recommendations could form the basis of an ES.

Financial Advisor – Asset Allocation

Investment firms often provide brochures or pamphlets to help make financial decisions such as asset allocation. For example, based on the investor's age, annual income, net worth, risk tolerance, and anticipated retirement suggestions are presented as to how the investment might be allocated between investment categories. The categories usually include Income, Growth, and combinations of the two. An ES could be constructed from such materials.

Blackjack Advisor

Develop a system to advise you whether to "hit" or "stay" based on your hand and what the dealer has showing. If the suggestion is "hit" then the system should advise whether to "double down." Rules for such decisions can be found in most books on Blackjack. If you choose this project and cannot find an appropriate book or pamphlet I have one.

Course Advisor

Develop a system to advise students on course selection. This can be based on the Saint Ambrose course catalog for undergraduate or graduate students. Inputs include academic requirements (university, college and department), student status (what he/she has already taken) and personal preferences. Its outputs will suggest courses to take next. Do not assume that a student has taken all the courses listed on the standard schedule sheet through his or her prior semester, or recommend the block schedule from that schedule sheet as it stands. Since this is a decision support system, not a decision-making system, it does not have to specify each course precisely. It can leave some choices (e.g., electives) to its user.

Personal Ad Advisor

A possible project, which I have never seen done but which I suspect would be fun, would be deciding which 'personals' ad to answer. Do I go for the one who really wants someone a bit younger (or older) than I am, but who likes the same sports and music that I like, or the one who is perfect age-wise but doesn't like beards? How should I deal with missing information? This is not the same as a date-matching program, as all participants in date-matching services fill out the same structured questionnaire and their answers can therefore be compared directly with what other participants are looking for. Personal ads are more free form. However, the scope of this project is more limited than a general computer-dating program: it need consider only the needs of one person, not all the potential needs of the human race. The use of negative information (potential incompatibilities, such as her distaste for beards) would complicate this project in EXSYS but would be less of a problem with a more capable shell. Since a person can answer more than one ad, the ideal output would give several top candidates with numerical ratings.

Consumer Advisor

Produce an expert system, which will help a consumer to choose an electrical product of some sort. You are free to choose the product you want to deal with: CD players, microwaves, TV's, vacuum cleaners, power tools, digital cameras, camcorders, scanners would all be reasonable choices. You want a relatively complex product, with a range of features (price, size, weight, color, levels of sophistication of various sorts) which makes choosing a specific model a reasonably difficult task for the layperson. You should aim at including around 15-20 example products.

Wine Advisor Copyright © 1996 Royal Melbourne Institution of Technology

Develop an expert system to select a wine to serve with certain dishes in various situations. Below is some information for you to encode as part of the expert system. (I'm no expert on wine selection so feel free to embellish the information if it offends you :-)

Consider wines which accompany dishes (as opposed to other drinks). Important properties of a wine include its color, sugar and grape-type. Wines are also classified by their region. We'll make it easier and only consider the country the wine comes from: we'll limit attention to French--and only the Bordeaux region--and Australian wines.

Some grape types: chardonnay, produces a white wine; pinot-noir, produces a red; cab-sav and merlot, both produce reds, with sugar-type dry; two types of riesling: riesling is white, but can be sweet or dry sugar-type. Bordeaux wines are typically red, but may be white; red bordeaux's are made from either cab-sav or merlot grapes, while white bordeaux's are made from semillon or sav-blanc grapes.

There are all sorts of dishes that you are capable of preparing: seafood, including fish or shellfish; pasta dishes with different sorts of sauces; chicken dishes; and various red-meat dishes. All of these dishes may be spicy or bland. You can also prepare some basic snacks, like a fruit platter, or a cheese platter.

For a fruit platter, an appropriate wine is a sweet white wine, whereas for a cheese platter you should serve a red wine. For seafood, white wine is appropriate; for fish, the wine should be dry. For red meat, you should serve red; for other meat (including, but not only, chicken), you should serve white, and for chicken, it should be dry. A red wine should accompany pasta with red sauce, while a white wine should accompany pasta with white sauce; this even if the red-sauce contains chicken and the white sauce contains red meat.

Being a cheapskate, you're going to serve to serve a poor-quality wine unless you have a visitor who you are trying to impress: then you would serve a good-quality wine. People who you may want to impress include your boss and your boyfriend or girlfriend (choose one ...). For any visitor, you will also try to match any preference they may have over and above any other requirement.

Some specific wines: CloudyBay produces a good quality Australian chardonnay, so does Penfolds. You can get a good Aust. semillon from Yalumba and a good Aust. reisling (both dry and sweet varieties) from Eaglehawk, and Lilly Pilly Estate produce a good sweet semillon. De Bortoli (Aust) produce a good cab sav, Yarra Ridge a good pinot noir; Piero produces good wines of both varieties.

For poor-quality Australian wines, you can go for a Lindeman cask wine (sweet or dry rieslings, or generic 'red'), Queen Adelaide cab sav, Matthew Lang chardonnay, or Oxford Landing 'red'.

There's no point importing poor-quality French wines, so anything French you serve will be good. Chateau Cheval makes white Bordeaux' (of all varieties) and Chateau d'Ychem produces all sorts of Bordeaux reds.

Naming wines: Wines should be named by combining the name of the wine manufacturer with the grape-type (if known--otherwise just use the color). For example: cloudybay-chardonnay, penfolds-chardonnay, lindeman-dry-riesling, lindeman-red, etc. This will assign unique names to the wines above.

Some specific dishes you know how to cook: steamed lobster, barbecued prawns (both shellfish); grilled swordfish; spicy fried chicken; roast beef; duck a l'orange; beef curry (spicy!); fettuccine al fredo (with a cream sauce); tortellini gorgonzola (white cheesy sauce); pasta diavolo (spicy tomato-based sauce); spaghetti bolognese (spicy red meat sauce). You can also prepare cheese and crackers, and a fruit and nut platter.

If you are feeling rich, then you would serve good wine, no matter who was visiting. You would be feeling rich if you won Tattslotto, or if you won at the casino.

Test scenarios: