Chapter 4

 

Decision Support and Artificial Intelligence

 

Brainpower for Your Business

 

Chapter Map

Introduction

n     People make decisions all the time.

n     In a business environment, managers regularly make complex decisions.

Introduction

Management Review à

Future successes will be for those organizations that are

Big on brain and small of mass”.

Introduction

The categories of computer-aided decision support.

 

Decisions, Decisions, Decisions

Decisions, Decisions, Decisions

 

 

Decisions, Decisions, Decisions

 

 

Decisions, Decisions, Decisions
How You Make a Decision

 

 

Decisions, Decisions, Decisions
How You Make A Decision

Step 1: Intelligence

Find or recognize a problem, need, or opportunity.

n    This is the diagnostic phase of the decision making

n    Involves the detection and interpretation of signs

t   Consistent customer request for new feature or product

t   Threat of new competition

t   Declining sales

t   Rising costs

t   Offer from a company to handle your distribution needs

Decisions, Decisions, Decisions
How You Make A Decision

Step 2: Design

Consider ways to solve the problem, fill the need, or take advantage of an opportunity.

 

This is the phase where you would study all possible solutions and develop those that are most interesting and feasible.

 

Decisions, Decisions, Decisions
How You Make A Decision

Step 3: Choice

Examine and weigh the merits of each solution, estimate the consequences of each, and choose the best one.

 

n     This is the prescriptive phase of the decision making.

n     In this stage a course of action is prescribed and this course of action could simply be to do nothing.

n     Selection of best solution requires selection criteria such as:

t    Cost

t    Ease of implementation

t    Staffing requirements

t    Timing

Decisions, Decisions, Decisions
How You Make A Decision

Step 4: Implementation

Carry out the chosen solution, monitor the results, and make adjustments as necessary.

 

The implementation of final solution is seldom enough and will practically always need fine-tuning.

 

Decisions, Decisions, Decisions
Types of Decisions You Face

Figure 4.3
Viewing Structured Versus Nonstructured Decision Making as a Continuum
page 136

Decisions, Decisions, Decisions
Types of Decisions You Face

n      If you want to decide which cheese to by then the criteria would be based on:

t     Expiry date

t     Price

n      If you have to decide which job offer is better to accept, then the criteria for your decision would be based on:

t     Salary

t     Intrinsic motivation

t     Location

t     Company

t     Position

t     Benefits

t     Many other criteria

Decisions, Decisions, Decisions
Types of Decisions You Face

n     In the cheese case, the criteria are pretty straight forward, hence the decision making is rather STRUCTURED.

 

n     In the case of the job offer, the criteria are many and most are not straight forward and are difficult to quantify, hence UN-STRUCTURED.

Decisions, Decisions, Decisions
Types of Decisions You Face

Structured decision

Processes information in a specified way so that you will always get the right answer.

n     Does not involve any “feel” into the process

n     Does not require intuition

n     These types of decision can be easily programmed

t    Fixed set of input

t    Same processing

t    Produce correct result

Decisions, Decisions, Decisions
Types of Decisions You Face

Non-structured decision

One for which there may be several “right” answers and there is no precise way to get a right answer.

n     No rules or criteria exist that guarantee you a good solution.

n     Examples

t    To introduce a new product line or not

t    To employ a new marketing campaign

t    To change the corporate image

Decisions, Decisions, Decisions
Types of Decisions You Face

Recurring decision

One that happens repeatedly, and often periodically.

 

n     Usually use the same set of rules each time.

Decisions, Decisions, Decisions
Types of Decisions You Face

Nonrecurring decision

One that you make infrequently.

n     Or ad-hoc decision

n     May have different criteria for determining the best solution each time

n     Example

t    Company merger

Decision Support Systems

Decision Support Systems

Highly flexible and interactive IT system that is designed to support decision making when the problem is

not structured.

 

Decision Support Systems

n    A DSS is an alliance between

t   you,

t   the decision maker and

t   specialized support provided by IT

Decision Support Systems

n     A DSS brings

t    Speed

t    Vast amounts of information

t    Sophisticated processing capabilities

Decision Support Systems

n     You bring the know-how in the form of

t    Experience

t    Intuition

t    Judgment

t    Knowledge of relevant factors

n     You must know the right kind of questions to ask and how to process the information so you may get useful answers from the IT.

Decision Support Systems

Figure 4.4
The Alliance Between You and a Decision Support System
page 137

 

Decision Support Systems
Components of a Decision Support System

n    A typical DSS has three components:

t   Model management

t   Data management

t   User interface management

Decision Support Systems
Components of a Decision Support System

Figure 4.5
Components of a Decision Support System    page 138

 

Decision Support Systems
Components of a Decision Support System

How do the components work together?

n      You tell the DSS using the user interface management component which model to use

n      The model is found in the model management component which you had instructed to use on a set of information found in the data management component

n      So the model

t     Requests the information from the data management component

t     Analyzes that information

t     Sends the results to the user interface management component, hence back to you

 

Decision Support Systems
Components of a Decision Support System

The Model management

Consists of both the DSS models and the DSS model management system.

t    A model is a representation of some event, fact or situation to represent variables and their relationships

t    The model you use in DSS depend on the decision you are making and consequently the kind of analysis you require such as:

"    What-if analysis

"    Optimization

 

Decision Support Systems
Components of a Decision Support System

Data management

Performs the function of storing and maintaining the information that you want your DSS to use.

n      It consists of both the DSS information and the DSS database management system

n      The information you use in your DSS comes from one or more of the three sources:

t     Organizational information

t     External information

t     Personal information

Decision Support Systems
Components of a Decision Support System

User interface management

Allows you to communicate with the DSS.

n      Consists of the user interface and the user interface management system

n      Allows you to combine your know-how with the storage and processing capabilities of the computer

n      Through it you enter

t     Information

t     Commands

t     Models

n      Best user interface uses your terminology and methods and is:

t     Flexible

t     Consistent

t     Simple

t     adaptable

Collaboration Systems

 

Collaboration system

 

A system that is designed specifically to improve the performance of teams by supporting the sharing and flow of information.

 

 

Figure 4.6
Collaboration Software
Connects People
page 141

 

Collaboration System

Collaboration software takes many forms with

t   Many combinations of features

t   Varying degrees of complexity

 

 

Collaboration System

Three types of collaboration systems are discussed in the book:

t   Enterprise-wide collaboration

t   Supply chain collaboration

t   Web-based collaboration

 

 

Collaboration System
Enterprise-wide Collaboration

n    Lotus Notes and Microsoft Exchange are types of integrated collaboration systems.

 

 

Collaboration System
Supply-Chain Collaboration

n    Supply chain management means working with your suppliers and distributors in all phases of planning, production, and distribution.

 

 

Collaboration System
Web-Based Collaboration

n     Web-based collaboration tools use the power of the Internet to enable people to work together effectively and efficiently.

 

n     The peer-to-peer file-sharing feature is combined with the ability to create and edit documents collaboratively, and to send and receive text and voice messages.

Geographic Information Systems

Geographic Information Systems

Geographic information system (GIS)

 

A decision support system designed to work with spatial information.

Geographic Information Systems

Spatial information

 

Is any information that can be shown in map form, such as roads, the distribution of bald eagle populations, and the layout of electrical lines.

Geographic Information Systems

Figure 4.7
Geographic Information Systems
page 144

Geographic Information Systems

n    Today GISs are helping businesses:

t   Identify the best site to locate a branch office

t   Target pockets of potential customers

t   Reposition promotions and advertising

t   Determine the optimal location of a new distribution outlet

Geographic Information Systems

Business Geography

 

A new type of information created when businesses combine textual information with spatial information

Artificial Intelligence

Artificial Intelligence

Artificial Intelligence

Artificial intelligence (AI)

 

The science of making machines imitate human thinking and behavior.

 

 

Artificial Intelligence

Robot

 

A mechanical device equipped with simulated human senses and the capability of taking action on its own.

 

 

Artificial Intelligence

n    Who uses artificial intelligence?

t    Financial analysts to:

"    Manage assets

"    Invest in the stock market

"    Perform other financial operations

Artificial Intelligence

n    Who uses artificial intelligence?

t    Hospitals:

"    For scheduling staff

"    Assigning beds to patients

"    Diagnose illnesses

"    Treat conditions and illnesses

Artificial Intelligence

n    Who uses artificial intelligence?

t    Credit card companies

"    To detect credit card fraud

t    Insurance companies

"    To ferret out fraudulent claims

Artificial Intelligence

n     Artificial intelligent systems that businesses today use most can be classified into the following major categories

t    Expert systems

t    Neural networks

t    Genetic algorithms

t    Intelligent agents

Expert Systems

Expert Systems

Expert system

or knowledge-based system

 

Is an artificial intelligence system that applies reasoning capabilities to reach a conclusion.

Expert Systems
Components of an Expert System

n     Expert systems are good for

t    Diagnostic problems: Problems requiring an answer to the question “what’s wrong?” correspond to the intelligence phase of decision making and

t    Prescriptive problems: Problems that require an answer to the question “what to do?” and correspond to the choice phase of decision making.

Expert Systems
Components of an Expert System

Figure 4.9
Traffic-Light Expert System Rules
page 149

Expert Systems
Components of an Expert System

n     An expert system is usually built for a specific application area called domain:

t    Accounting

t    Medicine

t    Process control

t    Human resource management

t    Financial management

t    Production

t    Forestry management

 

Expert Systems
Components of an Expert System

n    An expert system combines

t   Information types,

t   People, and

t   IT components.

 

 

Expert Systems
Components of an Expert System

n    Information types include:

t   Domain expertise

"   The reasoning process that will solve the problem.

t   “Why” information.

t   Problem facts.

 

 

Expert Systems
Components of an Expert System

Figure 4.8
Developing and Using an Expert System
page 150

Expert Systems
Components of an Expert System

n     People

t    Domain expert

"    provides the domain expertise in the form of problem-solving strategies.

t    Knowledge engineer

"    IT specialist who formulates the domain expertise into an expert system.

t    Knowledge worker or user

"    that’s you.

Expert Systems
Components of an Expert System

n     IT Components

t    Knowledge base

"    stores the rules of the expert system.

t    Knowledge acquisition

"    used to enter the rules.

t    Inference engine

"    takes the problem facts and searches the knowledge base for rules that fit.

t    User interface

"    used to run the consultation.

Expert Systems
What Expert Systems Can and Can’t Do

n     An expert system can:

t    Reduce errors

t    Improve customer service

t    Reduce costs

 

n     An expert system can’t:

t    Use common sense

t    Automate all processes

Neural Networks

Neural Networks

Neural network

(often called an artificial neural network or ANN)

 

an artificial intelligence system that is capable of finding and differentiating patterns.

Neural Networks

n     A NN simulates the human ability to classify things without taking prescribed  steps leading to the solution

n     A NN can learn by example

n     A NN can adapt to new concepts and knowledge

n     NN are widely used for visual pattern and speech recognition systems

n     A PDA probably uses NN to decipher your handwriting

 

Neural Networks

n     NN are used for many other applications such as

t    In airports to detect bombs

t    In police departments to identify corruption

t    In medicine to

"    Check for drug interaction

"    Detect anomalies in tissue samples

"    Heart attacks

 

Neural Networks

n    In business NN are used

t   Security trading

t   Fraud detection

t   Real estate appraisal

t   Evaluating loan applications

t   Target marketing

 

Neural Networks
Types of Neural Networks

n     Self-organizing neural network - finds patterns and relationships in vast amounts of data by itself.

 

n     Back-propagation neural network - a neural network trained by someone.

Neural Networks
Inside a Neural Network

Figure 4.11
The Layers of a Neural Network
page 155

Neural Networks
 
Inside a Neural Network

n    NNs attempt to mimic the human brain

n    They consist of three layers of a virtual neuron

t   Input layer

t   Hidden layer

t   Output layer

 

Neural Networks
 
Inside a Neural Network

n     Advantages of a neural network

t    Can learn and adjust to new circumstances on their own

t    Lend themselves to massive parallel processing

t    Can function without complete and well structured information

t    Can cope with huge volumes of information with many dependent variables

t    Can analyze nonlinear relationships in information

Neural Networks
 
Inside a Neural Network

n    The biggest problem with NN

t   The hidden layers are hidden

Genetic Algorithms

Genetic Algorithms

Genetic algorithm

 

An artificial intelligence system that mimics the evolutionary, survival-of-the-fittest process to generate increasingly better solutions to a problem.

 

Genetic Algorithms

n     Genetic algorithms are artificial intelligence software capable of following trial-and-error processes leading to the evolution of a good result.

n     In other words, genetic algorithms are optimizing systems such that they find the combination of inputs that give the best outputs.

Genetic Algorithms

n     Genetic algorithms use three concepts of evolution:

t    Selection — survival of the fittest.

t    Crossover combining portions of good outcomes in the hope of creating an even better outcome.

t    Mutation randomly trying combinations and evaluating the success (or failure) of the outcome.

Genetic Algorithms

n     Who uses genetic algorithms?

t    They are used by business executives to help them decide which combination of projects a firm should invest in.

t    They are used by investment firms to help then in trading choices and associated decisions.

t    Used in the garment industry to help them solve the problem of laying out the pieces of the garment and cutting fabric in a way that leaves as little waste as possible.

t    Used to determine the optimal configuration of fiber optic cable in a network that may include as many as 100,000 connection points.

Genetic Algorithms

Intelligent Agents

Intelligent Agents

Intelligent agent

 

Software that assists you, or acts on your behalf, in performing repetitive computer-related tasks.

Intelligent Agents

n    Four types of intelligent agents include:

t    Buyer agents or shopping bots

t    User or personal agents

t    Monitoring-and-surveillance or predictive agents

t    Data-mining agents

Intelligent Agents
Buyer Agents

Buyer agent or shopping bot

an intelligent agent on a Web site that helps the customer find products and services.

 

Intelligent Agents
Buyer Agents

n    Types of filtering include:

t   Collaborative filtering - a method of placing you in an affinity group of people with the same characteristics.

 

t   Profile filtering - requires that you choose terms or enter keywords.

Intelligent Agents
Buyer Agents

n     Types of filtering continued:

t    Psychographic filtering - anticipates your preferences based on the answers you give to a questionnaire.

 

t    Adaptive filtering - asks you to rate products or situations and also monitors your actions over time to find out what you like and dislike.

 

Intelligent Agents
User Agents

n    User agents (sometimes called personal agents) - intelligent agents that take action on your behalf.

Intelligent Agents
Monitoring-and-Surveillance Agents

n    Monitoring-and-surveillance agents (also called predictive agents) - observe and report on equipment.

 

 

 

Intelligent Agents
Data-Mining Agents

n    Data-mining agent - operates in a data warehouse discovering information.

 

Intelligent Agents
Components of an Intelligent Agent

n     Autonomy - act without your telling them every step to take.

 

n     Adaptivity - discovering, learning, and taking action independently.

 

n     Sociability - conferring with other agents.

Closing Case Study One
Using Neural Networks To Categorize People

n     Using neural network software, businesses now have the ability to look for patterns in their customer information.

 

n     How accurate is it for a business to predict the future behavior of customers on the basis of past behavior?

 

Closing Case Study Two
Decision Support and Artificial Intelligence in Health Care

n     Good health care is based largely on good information.

 

n     How can DSS and AI be used to track symptoms, treatment, and outcomes that require the collection and maintenance of a huge amount of qualitative and quantitative information?

 

 

Summary
Student Learning Outcomes

n  Define decision support system, list its components, and understand its applications.

n  Define collaboration systems along with their features and uses.

n  Define geographic information systems and state how they differ from other decision support tools.

 

Summary
Student Learning Outcomes

n Define artificial intelligence and list the different types that are used in business.

n Define expert systems, and the type of problems to which they are applicable.

n Define neural networks, their uses, and the major strength and weaknesses.

Summary
Student Learning Outcomes

n Define genetic algorithms and list the concepts on which they are based, and the types of problems they solve.

n Define intelligent agents, list the four types, and identify the types of problems they solve.

 

 

Summary
Assignments & Exercises

n Make a GIS

n Collaboration Work

n Choose a Financing Option

n Which Software Would You Use?

Real Hot Electronic Commerce
Finding Investment Opportunities on the Internet

n     Learning about Investing

n     Researching the Company behind the Stock

n     Finding other Sources of Company Financials

n     Making Trades Online

n     Retrieving Stock Quotes

 

Visit the Web to Learn More
www.mcgrawhill.ca/college/haag

n     Learning about Investing

n     Researching the Company behind the Stock

n     Finding other Sources of Company Financials

n     Making Trades Online

n     Retrieving Stock Quotes