Transcript SOM485CH3CLASSSLIDES

Chapter 3
Study sub-sections: 3.5-10, 3.12(p118-120)
DECISION SUPPORT
SYSTEMS
CONCEPTS,
METHODOLOGIES,
AND
TECHNOLOGIES:
AN OVERVIEW
Components of a Closed DSS
Components of an Open DSS
Network
DSS Classifications
(based on dominance of each component)
– Model-driven DSS: quantitative models (statistical, financial,
optimization, simulation) used to generate a recommended
solution to a problem
– Data-driven DSS: support ad-hoc reporting and queries on
internal & external database
– Communication-driven: multiple user interface, support shared
tasks, either cooperative or hostile mode
– Knowledge-driven: qualitative models; uses stored rules
(Expert Sys & Mining)
– Document-driven: search, retrieve, analyze, classify text
documents (eg. Law firms use it to create a case)
Data Management Subsystem-1
Data Management Subsystem-2
• The Database
– Internal data come mainly from the organization’s
transaction processing system
– External data include industry data, market research data,
census data, regional employment data, government
regulations, tax rate schedules, and national economic
data
– Private data can include guidelines used by specific
decision makers and assessments of specific data and/or
situations
Data Management Subsystem-3
• Data extraction
The process of capturing data from several disparate
sources, synthesizing them, summarizing them, determining
which of them are relevant, and organizing them, resulting in
their effective integration
Data Management Subsystem-4
• Database management system (DBMS)
• Software for establishing, updating, and querying a database
• Directory
A catalog of all the meta-data in a database or all the models
in a model base
Data Management Subsystem-5
• Key database and database management system
issues
–
–
–
–
Data quality
Data integration
Scalability
Data security
The Model Management Subsystem-1
The Model Management Subsystem-2
• Model Directory
• Index/Catalog/List of all models (meta-data on models)
• Model Base
Contains the actual collection of available models themselves
that can be readily instantiated with data
• Model Base Management
• Tools for creating, manipulating, updating
The Model Management Subsystem-3
• Four categories of models in the model base
(based on Business Functions)
–
–
–
–
Strategic models
Tactical models
Operational models
Analytical models
Model integration involves combining the operations of
several models when needed; Eg. Factor analysis
determines which variables are most promising and
regression analysis follows up with creating the actual
prediction model.
The Model Management Subsystem-4
• Strategic models
Models that represent problems for the executive level of management (eg. How
many plants should we have five years from now? Uses considerable external
data)
• Tactical models
Models that represent problems for the mid-level of management (eg. Short-term
labor recruitment&training, sales promotion planning, budgeting)
• Operational models
Models that represent problems for the operational (day-to-day activities) level of
management (eg. Production scheduling, staffing, inventory control)
• Analytical models
Mathematical models typically integrated into the above models
Egs.Statistical, Financial, MS, data mining algorithms
The Model Management Subsystem-5
• Model building blocks /routines
– Helps to create a custom model from smaller components
– Preprogrammed software elements that can be used to
build computerized models.
For example, a random-number generator can be
employed in the construction of a simulation model
– Models created using blocks /routines can easily be
updated
– Some programming is required
– Modeling languages (MDX, XMLA similar to SQL for DBs)
can also be used
User Interface (Dialog) Subsystem-1
• User interface management system (UIMS)
The DSS component that handles all interaction between users and the
system
+GUI
User Interface (Dialog) Subsystem-2
• Since managers are used to verbal interactions and have
time constraints, designing DSS interfaces pose a
challenge
– Voice input and output (no typing)
– Natural language processing (typing spoken language)
– GUI (more info can be presented compared to text)
– Touchscreen (slice & dice data)
– Responding to body movements (including face)
– Portable devices/ Web interface (as they travel a lot)
– Intelligent agents & search engines
– Flexibility to suit styles of decision-maker
– Ability to support group decision-making
Knowledge-Based Management Subsystem
(Chapters 11-12)
•
•
•
•
•
•
•
•
Optional component of a DSS
Typically captures qualitative knowledge mathematical models
Static knowledge of a domain
Decision rules used in the domain (If-Then)
Heuristic / logical reasoning
Backward (goal to data)/ Forward chaining (data to goal)
Repository of past decisions and outcomes (machine learning)
Ability to consult other experts in the field
The User
A DSS may be directly used by a decision-maker.
But many managers employ an Intermediary/ chauffer:
A person who uses a DSS to answer questions for top
management
The intermediary should be a(n):
Expert tool user : with skills in the application of one or
more types of specialized problem-solving tools
Facilitator: who can plan, organize, and electronically
control a group in a collaborative computing
environment