ACS 1803Lecture Outline Decision Support

Systems

Decision Support Systems *L

-designed to help management make semi- structured (or unstructured) decisions

- but such systems do not make decisions (why?)

- they focus more on what might happen rather than what has happened

 

- typically include: *L

a) a data base, perhaps a "data warehouse", extracted from a "live“ database,

b) a model base*** that uses the data base [a model is a structured representation of some aspect of reality; it is because of the model that we can examine effects of decisions; but, a model always has assumptions e.g., inflation rate, net earnings level over 5 years; cost increases]

c) a user-friendly interface (dialog), often involving graphics 

Three Fundamental DSS Components

-a DSS may be developed by people outside of the Information Systems Department

- a DSS also can have capability for ad hoc reporting from the data base (warehouse)

examples of decision support: *L

- should we buy out a company? should we expand into another product line? [why semi-structured?I]

- classic illustration: Houston Oil and Mineral Co

Houston Oil and Minerals Example

• HOM was interested in a proposed joint venture but required a risk analysis

• They built a DSS using a specialized planning language (4GL named IFPS)

• Results from DSS model suggested the project be accepted

• The Executive VP using his intuition, experience and judgment asked that yet another possible but improbable condition be examined by the DSS

Houston Oil and Minerals Example ct’d

• The DSS was flexible and responsive enough to allow managerial judgment and intuition to be incorporated into the analysis within less than 1 hour

• Using the results that came back, the executive reversed the decision and rejected the joint venture proposal

• Ultimately this proved to be the right decision!

DSS Examples *L

• A more primitive example of a DSS is a spreadsheet used for “what-if” analysis

• There are Excel templates built for certain types of decisions [terms: template, model; explain these]

• A more complex DSS is a simulation model of a hospital’s surgical scheduling or other health care structures (as in You Tube videos)

Also see paper handout

for DSS examples

Model-Driven Ex. – Loan Calculator

10

Variables to be Analyzed Loan Calculator Model

Analysis Results

Medical Clinic Simulation Model• HC-Simulation Software to Optimize Healthcare

Processes - https://www.youtube.com/watch?v=neBCg7N1UyM

• Flexsim Heathcare Urgent Care Tutorial – Video 1– Video 2– Video 3

11

More DSS Examples

• Airline industry: DSS helps to find proper pricing to maximize overall revenue from selling seats for each flight– Mgr enters depart. pt, arr. pt, no of stops, times of

dep and arr, # days in advance for res, # persons, size of plane, utilized capacity on similar previous flights etc.

– System suggests variable ticket prices

DSS development and use

• Many DSS are not developed by computer professionals (at least not alone)– E.g., power sales support system at Manitoba

Hydro (engineer with MBA degree uses IFPS system in the Finance department)

• DSS are used largely by middle and higher managers

A Comparison of DSS and MIS

• DSS differs from an MIS in numerous ways, including: – The type of problems solved– The support given to users– The decision emphasis and approach– The type, speed, output, and development of the

system used– See comparison of DSS with MIS p. 292

14

Expert Systems *MC• p. 328-334• such systems are different than traditional reporting

or DSS systems

• they apply artificial intelligence to situations where many facts and complex decision rules are involved, such that only a few people can solve such problems well

• an expert system mimics the thinking of an expert

Expert Systems• Expert system manipulate knowledge and not

just information• e.g what drug and in what dose to give for

particular types of cancer– Many factors involved– Many questions must be asked– Many IF … THEN rules

• A rule is a way of encoding knowledge- an ES should be able to explain its reasoning to

the user

Expert Systems

• ***why develop them? *L- to retain expert's knowledge if he retires or dies- to pool expertise from several experts- to clone the expert's knowledge and have it available in many places at once (e.g., cancer treatment in remote Manitoba areas)

• they can be developed through detailed programming or through an "expert system shell" such as VP Expert

Expert System structure *L

• Knowledge base– Facts and rules

• Inference engine– Software that takes user input and “sifts through”

the knowledge base mimicking the mind of an expert

– See paper handout eg. of ES program• This is artificial intelligence

Expert System Development *MC

• A knowledge engineer has special expertise in eliciting information and expertise from experts

• He / she translates the expert’s knowledge into a set of (if .. then) rules

Expert Systems Examples *MC

• ES at California State U to advise students on class selection

• Complex machine repair• Cancer treatment in remote areas• Computer user help desk• See paper handout for HR example

Knowledge Management

• An expert system works on a knowledge base– It is part of a larger area called ‘knowledge

management’

Knowledge Management Definitions *MC

Knowledge AssetsAll underlying skills routines, practices, principles, formulae, methods, heuristics, and intuitions whether explicit or tacit

Tacit KnowledgeThe processes and procedures on how to effectively perform a particular task stored in a person’s mind

Explicit KnowledgeAnything that can be documented, archived, or codified often with the help of information systems

Knowledge ManagementThe process an organization uses to gain the greatest value from its knowledge assets

Knowledge Management System (KMS) *MC

Primary ObjectiveHow to recognize, generate, store, share, manage this tacit knowledge (Best Practices) for deployment and use

TechnologyGenerally not a single technology but rather a collection of tools that include communication technologies (e.g. e-mail, groupware, instant messaging), and information storage and retrieval systems (e.g. database management system) to meet the Primary Objective

Best PracticesProcedures and processes that are widely accepted as being among the most effective and/or efficient