IS5740: Management Support Systems

1
IS5740 Management Support Systems

• Fuzzy Logic and Hybrid Intelligent Systems
  (Sources Chapter18, Turban and Aronson,
  Precisely Fuzzy - Part I, Intelligent Software
  Strategies, Vol 9, No. 4, March1993)
• Fuzzy Logic FAQ

2
Fuzzy Logic Theory and Applications

• Fuzzy logic deals with imprecision and
  uncertainty
• Uses the mathematical theory of fuzzy sets
• Simulates the process of normal human reasoning
• Allows the computer to behave less precisely and
  logically

3
Fuzzy Logic Applications - DSS for Securities
Trading

• (source Chan and Bolloju, ISDSS '95)
• The Background
• Securities commercial papers, notes, bonds
• Qualitative and Quantitative nature of the
  expertise
• Approaches financial models, expert systems,
  artificial neural networks and fuzzy logic
• Suitability of fuzzy logic for knowledge
  representation

4
Architecture of the DSS
5
A Model for Buy Decisions
6
Relationships between the key variables
7
Resultant-Risk based on IssuerCR, BondCR and
CounterPartyCR
 
8
Support for Buy based on Yield-To-Maturity and
Resultant-Risk
   
9
Fuzzy Logic for Decision Modeling - Software
Effort Estimation

• (Source Bolloju 1993, An Expert System for
  Software Effort Estimation, Working paper)
•  
• Historical Data - not available or partially
  available
• System Characteristics - not known precisely and
  certainly

10
A Model for Estimation of Effort
11
A Model for Estimation of Effort

• Imprecision Uncertainty in Inputs and Process
•  System or subsystem characteristics such as
  numbers and complexities of entities, processing
  functions, reports, etc. are neither precise nor
  certain
• The process of estimation itself is vague and it
  is expected to provide good estimates using vague
  inputs

12
Concept of Imprecision and Uncertainty

•   of processing functions in a given subsystem

13
Conventional Sets
14
Fuzzy Sets and Membership Functions
15
Membership Functions

• User defined ( medium, large, tall, heavy, ...)
  and not just symbolic names
• Defined on a given domain (0-10, a,b,c,...,
  0-100)
• Map domain values to 0,1
• Used for specifying imprecise and / or uncertain
  values (medium complexity, heavy object, may be
  tall person, ...)
• Modifiers such as very, rather, ... can be
  applied (very tall, rather heavy, ...)
• Used for specifying fuzzy rules and fuzzy inputs
•  

16
Fuzzy Logic and Rules
17
Approximate Reasoning with Fuzzy Rules
18
Approximate Reasoning with Fuzzy Rules

•  Input parameters can be precise or imprecise
  and/or uncertain (e.g., complexity 8,
  complexity rather high)
• Rules are applicable to a degree between 0 and 1
• All rules applicable to the degree gt 0 (or some
  threshold such as 0.2) will get evaluated
• All the outputs produced by such rules are
  combined (defuzzification) to form the final
  output (e.g., centroid method)
• The final output can be treated as a precise
  value (if required)

19
Fuzzy Logic Advantages

• Provides flexibility
• Provides options
• Frees the imagination
• More forgiving
• Allows for observation
• Shortens system development time
• Increases the system's maintainability
• Uses less expensive hardware
• Handles control or decision-making problems not
  easily defined by mathematical models

20
Advantages

• Deals with both imprecision and uncertainty
• Lesser number of rules compared conventional
  production rules
• No need for complex mathematical modelling and /
  or simulation

21
Fuzzy Logic Applications and Software

• Used in consumer products that have sensors
• Air Conditioners
• Cameras
• Dishwashers
• Microwaves
• Toasters
• Special Software Packages like FuziCalc
  Spreadsheet
• Controls Applications
• Fuzzy TECH Home Page

22
Fuzzy Logic Applications

• Selecting stocks (on Japanese Nikkei Stock
  Exchange)
• Retrieving data (fuzzy logic can find data
  quickly)
• Regulating auto antilock braking systems
• Camera Auto-focusing
• Automating laundry machine operation
• Building environmental controls
• Controlling video camcorders image position
• Controlling train motion
• Identifying killer whale dialects

23
Fuzzy Logic Applications (contd.)

• Inspecting beverage cans for printing defects
• Keeping space shuttle vehicles in steady orbit
• Matching golf clubs to customer's swings
• Regulating shower head water temperature
• Controlling cement kiln oxygen levels
• Increasing industrial quality control application
  accuracy and speed
• Sorting multidimensional space problems
• Enhancing queuing (waiting lines) models
• Decision making (see Glenn 1994)

24
Applications

• Control Systems Transport, Transmission,
  Cameras, Washing machines, Air conditioners, ..
• Information Systems Finance (loan appraisal,
  investment analysis, stock trading, ...),
  Criminal Investigation, Risk assessment, ...

25
Cross Fertilization Hybrids of Cutting Edge
Technologies

• Combine
• Neural Computing
• Expert Systems
• Genetic Algorithms
• Fuzzy Logic

26
Summary

• Fuzzy logic represents uncertainty by using fuzzy
  sets
• Fuzzy logic is based on 1) People reason using
  vague terms. Classes boundaries are vague and
  subject to interpretation 2) Human
  quantification is often fuzzy
• Fuzzy sets have well defined boundaries. Items
  have membership values to define the imprecise
  nature of belonging to a set