Artificial Intelligence in Business
450.526

Unit Outline 2003


 

Unit Description

The use of computational intelligence in management, business and administration. Enough attention is given to the theoretical bases for a working understanding but the focus is on applying available computer packages.

Objectives

The aim is to understand and apply expert systems, genetic algorithm, fuzzy logic and neural networks in a business management and planning environment. The course is designed to enable students to:

• Achieve some understanding of applied computer based systems which learn to solve problems - artificial intelligence and evolutionary computing

• Develop a 'hands-on' capacity to apply these methods

• Assess the potential to apply artificial intelligence, evolutionary computing and expert systems in business and administrative contexts

• Judge where to use these methods and the strengths of each

• Compare artificial intelligence and evolutionary computing with analytical and simulation approaches to solving business problems

• Interpret and judge the work of specialised analysts

LECTURER

Professor John Taplin

Office:

Social Science South 2219

Telephone:

380 2081

Fax:

380 1004

E-mail:

jtaplin@ecel.uwa.edu.au

Class time

Monday 5.00-8.00 pm

Venue:

Social Sciences South (SSS) G.210 and Computer Lab SSS 2202

READING

There is no prescribed text. Readings will be distributed from time to time. It is often useful to scan books and papers for the more comprehensible sections. 

Fuzzy Logic

Mancini, D., Squillante, M. and Ventre, A. (eds.) New Trends in Fuzzy Systems, World Scientific, 1995 (FIZ 006.3, 1998 NEW)

Sanchez, E., Shibata, T. and Zadeh, L.A. Genetic Algorithms and Fuzzy Logic Systems, World Scientific, 1995 (HSS 006.3, 1997 GEN)

Zimmermann, H.-J. Fuzzy Sets and Decision Analysis, TIMS/Studies in the Management Sciences,Vol.20, Elsevier North-Holland,1984

Artificial Neural Networks

Caudill,M. and Butler,C. Understanding Neural Networks: Computer Explorations, MIT Press, 1992 [a relatively simple text]

Kvaal, K. and Djupvik, H. (1996) "Prediction of Customer Segments with Neural Nets", Marketing Research Today, November

Patterson, D.W. Artificial Neural Networks, Prentice Hall, 1996 (FIZ 006.3, 1996 ART)

Phillips, P., Davies, F. and Moutinho, L (1999) "The Interactive Effects of Strategic Planning on Hotel Performance: A Neural Network Analysis", Management Decision, vol.37, no.3, pp.279-288

Roberts, S.J. and Penny, W. (1997) "Neural Networks: Friends or Foes", Sensor Review, vol.17, no.1, pp.64-70

Venugopal, V. and Baets, W. (1994) "Neural Networks and Statistical Techniques in Marketing Research", Marketing Intelligence & Planning, vol.12, no.7, pp.30-38

Genetic Algorithms

Gatarski, R. (1999) "Evolutionary Banners: An Experiment with Automated Advertising Design", COTIM-99

GECCO '99 (Banzhaf et al eds.) Proceedings of the Genetic and Evolutionary Computation Conference, Morgan Kaufmann, 1999

Goldberg, D.E. Genetic Algorithms, Addison-Wesley, 1989 [much of it is readable]

Hurley, S., Mountinho, L. and Stephens, N.M. (1995) "Solving Marketing Problems Using Genetic Algorithms", European Journal of Marketing, vol.29, no.4, pp.39-56

Michalewicz, Z.Genetic Algorithms + Data Structures = Evolution Programs, Springer-Verlag, 1992 [skim for principles and ideas]

Rawlins,G.J.E.(ed.) Foundations of Genetic Algorithms, Morgan Kaufmann, 1991

Whitley, L.D. (ed.) Foundations of Genetic Algorithms•2, Morgan Kaufmann, 1993

Expert and Hybrid Systems

Giarratano, J. and Riley, G. Expert Systems, 2nd edn, PWS Publishing, 1994

Gonzalez, E.L. and Fernandez, M.A.R. (2000) "Genetic Optimisation of a Fuzzy Distribution Model", International Journal of Physical Distribution & Logistics Management, vol.30, no.7/8, pp.681-696

Li, Shuliang (2000) "The Development of a Hybrid Intelligent System for Developing Market Strategy", Decision Support Systems, vol.27, pp.395-409

Preux, Ph. and Talbi, E-G. (1999) "Towards Hybrid Evolutionary Algorithms", International Transactions in Operational Research, vol.6, pp. 557-570

Many books and conference proceedings are available in the Mathematical and Physical Sciences library as well as the Science Reading Room on Neural Networks.  Most have call number 006.3.

COMPUTING MANUALS

Fuzzy

CubiCalc RTC, Version 2, HyperLogic Corporation

Neural Network

NeuroShell 2, Ward Systems, 1999 (no manual sent - rely on the 'Help')

Genetic Algorithm

Evolver 4.0 Professional Edition

ASSESSMENT

The course grade will comprise:

Assignment 1: Fuzzy  (due on Wednesday 27 August) 15%
Assignment 2: Neural Net  (due on Wednesday 17 September) 15%
Assignment 3: Genetic Algorithm (due Wednesday 15 October) 15%
Presentation  10%
End of semester examination  45%

Marks may be scaled according to Faculty policy

COMPUTING

Instruction in each technique will be given in the computing lab. This will normally occupy about two of the three class hours. No previous experience is needed. Software packages to be used in the course are listed in the following table and are installed in the computing lab.

PROGRAM

Week

Prior Reading

Discussion Topic

Computing

Software

1

Introduction to computational intelligence

2

Giarratano and Riley 283-324

Introduction to theory of fuzzy sets and fuzzy logic

Initial exploration of fuzzy software

CubiCalc

3

CubiCalc manual

Fuzzy for decision and multiple attributes

Applying fuzzy to business decisions

CubiCalc

4

Patterson 1-36, 113-140; Roberts & Penny

Introduction to artificial neural networks (ANN)

Initial exploration of ANN software

NeuroShell 

5

Kvaal & Djupvik; Venugopal & Baets; Phillips, Davies et al.

Advanced applications of artificial neural networks (ANN)

Applying ANN to forecasting, and assessment

NeuroShell 

6

Goldberg 1-25, 89-120 Michalewicz 13-31

Introduction to genetic algorithm (GA); crossover, mutation, 'fitness', convergence

Initial exploration of GA software

Evolver

7

GECCO '99: 
East 1510-1516;
Weiss 718-725

GA theory & application, coping with intractable problems, multiple optima

GA for decisions, optimal scheduling

Evolver

8

Hurley, Moutinho & Stephens

GA theory & application, multiple optima; GA for marketing

Apply GA to the supermarket location problem

Evolver 

9

Preux & Talbi; Kingdon, 133-153 (in Sanchez, Shibata et al.)

Knowledge and rule based expert systems

10

Intelligent systems for market startegy (handout) Hybrid intelligent system for market strategy (Li)

  • Briefly discuss Ass.2

  • Formulate a hybrid intelligent system for market strategy

Exact and near exact solutions

Use Solver to finalise an Evolver solution

Excel Solver & Evolver

11

Solving marketing optimisation problems using genetic algorithms (Hurley et al)

  • Briefly discuss GA for marketing
  • Set up hybrid intelligent system for market strategy in CubiQuick

Implement the hybrid intelligent system

CubiCalc

12

Evolutionary banners: an experiment...
Giarratano & Riley 316-319
Fuzzy sets and fuzzy logic (handout) pp.3,4

  • Evolutionary algorithms in marketing

  • Applying the max-min fuzzy operator

PRESENTATIONS

 

 

 

[pres. in lab]

Powerpoint

13

PRESENTATIONS

Preparation for the exam

[pres. in lab]

Powerpoint