Transcript Chapter 1

Management
Decision Support
Overview for Today
• Exam coming up next week…
• Draft of project due Fri Apr 10…
• Management & Decision Making
– Why decision making is hard
• Models to support decision making
• Technologies to support decision making
So Just What is a DSS?
• Interactive systems that help decision
makers use data and models to solve semi& un-structured problems
– A counter example: Towers of Hanoi
– What are examples of these types of problems?
• Used by middle & upper level management
since 1970’s for analytical & quantitative
decisions
Difficulty of Decision Making
• ALL managerial jobs
require decision
making skills
• People are BAD at
decision making (study
from Gettys)
Difficulty of Decision Making
The Need for Decision Support
• Impossible to make decisions without information
– What’s the relationship between quantity of information
and decision quality?
– Moore’s law also relevant…
• New pressures/needs for improved decision making
– What are these new pressures?
– What are new needs of managers today?
– What factors create these new pressures?
(From our environ, business partners, customers)
Modeling & Decision Making
• Modeling is one way to aid decision making
• Generic tool, can be implemented w/o IT
• The business case for modeling:
– Why do/should managers use models?
– What are some benefits of modeling?
– Are there any drawbacks to modeling?
Four Types of Models
• Mental: description of how you think about a
situation, includes beliefs, assumptions,
relationships, used to generate descriptions or
make predictions, ex: “better to promote old
workers than younger ones” - this is often a first
step in modeling
• Scale: least abstract, physical replica of system,
used for designing cars, planes, production lines,
Bermuda…
Four Types of Models
• Analog: behaves like real system, doesn’t look
like it, ex: org charts, topo maps, blueprints
Four Types of Models
• Mathematical: complex modeling of weather
patterns, economic trends, physical systems, etc
– Eisenhower vs. Stevenson, 1952
– Bush vs. Gore in 2000; 2004
– Airplane design
More Mathematical Models
Decision Support Technologies
• Information systems to support decision making:
–
–
–
–
Decision Support Systems
Group Decision Support Systems
Artificial Intelligence
Expert Systems
• Each provides a different set of tools to aid the
decision maker
• Each is geared to a specific situation/context
• Goal: match the appropriate tech with needs of the
situation
Characteristics of a DSS
• Typical Feature Set:
–
–
–
–
–
Often based on mathematical models (Mgmt 102)
Integrates human judgment & data warehousing
Is adaptable by the user over time
Models provide sensitivity analysis
Utilizes models to promote learning – especially when
you can’t “play” with reality… `
DSS Examples
• GM Healthcare program
– Spent $4.5B in 2002, includes 127 providers
– $1,360 per vehicle for healthcare & pensions
• Uses OLAP & DSS to help slow the growth of
premiums
–
–
–
–
Compares diagnosis with treatments
Appropriate drug choices, generics, etc.
Correlate absenteeism with medical conditions
Allows ROI calculations for health care providers…
• “Good cop, Bad cop” system in LA
Group Decision Support Systems
• Similar to DSS with additional functionality:
– Teams of people work together to solve complex
problems
– Teams can be real or “virtual”
– Provides tool set for group process such as: anonymity,
brainstorming, voting, ranking, categorizing, group
consensus
– Tools help keep group coordinated/focused on task
• Examples of uses include:
– Strategic planning
– Support groups for victims of AIDS, rape…
Artificial Intelligence
• Attempts to represent human thought
process with machines
– Able to sense data, process it, draw its own
conclusions, act on those conclusions
• Examples:
– Automatic categorization useful with GDSS
applications: UA AI Categorizer
– “The Turing Test” (Alan Turing 1912-1954)
– IBM’s Watson, Deep Blue, Deep Jr.
AI Characteristics
• Capabilities:
–
–
–
–
–
Learn from past experience
“Sense Making” ambiguous / contradictory information
Quickly and successfully responding to a new situation
Infer rules from observed data
Use reasoning, logic as opposed to numeric formulas…
• Benefits
– Increase speed & consistency of problem solving,
especially when data are incomplete, inconsistent
– Helps in handling information overload by summarizing
or interpreting information
Expert Systems
• Allows employees to make better, faster decisions
–
–
–
–
–
Mimic human experts
Contains expertise from multiple individuals
Can support or replace decision makers
Can explain its recommendations
“Software IDs visible symptoms” story
• Some drawbacks
–
–
–
–
Expertise is hard to extract from people
Have very narrow fields of application
Construction is costly
Can be lack of trust by end users
Expert Systems
• Dept of the Treasury detects money laundering
– Banks report cash xacts > $10,000 (200,000/week)
– Don’t have the staff to examine each xact
– ES detect suspicious transactions and changes in
patterns; 100 cases/year since 1993
• State Street Bank and Trust Company
– Audits daily and month-end data against corresp GL
account balance, highlighting exceptions
– Increased the productivity of the auditors as well as
the quality of error detection
Automation of Mgmt’s Job
• Decision making involves specific tasks that
can take a long time to perform
– Automation can potentially save time, increase
consistency, enable better decisions
– Can decision responsibilities be completely
automated? (top mgmt? middle mgmt? lower
mgmt?)
– Decisions/tasks that can’t be automated?
– What are advantages? Disadvantages?
Important Concepts to Know
• Management & Decision Making
– Why decision making is hard
• Models to aid decision making
– Mental, Analog, Scale, Mathematical Models
• Types of Decision Support Technologies
– DSS, GDSS, AI, ES
For Next Time…
• No cases this time around
• Exam Review on Monday
–
–
–
–
–
Go through your notes
Re-do case studies
Make questions
Ask during class
Re-read “Exam Study Hints” on webpage