Decision Support System Dr. Ir. Sudaryanto, MSc. THE ARCHITECTURE FOR DECISION SUPPORT APPLICATIONS Part 2 Guidelines for Designing EIS Interfaces Dr.

Dr. Ir. Sudaryanto, MSc.Dr. Ir. Sudaryanto, MSc. Decision Support System

THE ARCHITECTURE FOR DECISION SUPPORT APPLICATIONS

Part 2Guidelines for Designing EIS Interfaces

Dr. Ir. Sudaryanto, MSc.Ir. (IPB), MSc (AIT-Thailand), Dr. rer.pol. (Aachen-Germany)


• Challenges in building EIS• Must be addressed on technical, organizational and • managerial issue• How to create EIS that easy to use• Attention on designing user interface

• EIS user interface• How the user direct the operation the system

• keyboard, mouse, touch screen, question/answer, command-• language, menu, voice)

• How the output is given to the user• graphical, textual, tabular, color, paper, online

• Due to the nature of the user• the system should be user friendly and user intuitive• the system should flexible

Development of EIS


Designing an EIS:

• Involve executives in design of the user interface• Setting standard for screen layout, format, color• Use the system should be intuitive• Use standard definition of terms• Design the main menu as a gateway to all computer use• Design the system for ease of navigation• Strive to make response times as fast as possible• Expect preferences in user interface to change

Design Guidelines


Involve executives in design of the user interface • Challenge: limited experience working directly with computer• Important to involve executive users in the process• Prototyping approach

• A full screen interface with large buttons and icons• A multiple windows interface with pull-down menus and • dialog boxes

• Multiple windows conform with interface design standard• Most of commercial EIS run under windows• Must be tailored according to the “computer literacy”• of executive users and their preference• Executive involvement in screen content and design• Option: train in windows and its application or • design application without windows• Must compromise between different requirements

Design GuidelinesThe Considerations


Setting standard for screen layout, format, color• Layout Standards

• graphical: visual presentation, • tabular: specific number• textual: explanation, assessment, action

• Graph standard• type of graphics : bar chart, line graphs, pie chart• avoid vertical wording

• Color standards• different color indicates different meaning• Traffic light-pattern for comparison, indicator

• Combination of layout, format and color consistent look and feel• reduce misinterpretation and misunderstandingUse of text• To make the information displayed more usefulAdvance capabilities• Voice commentaries, animation



Using the System Should Be Intuitive• Ideally, executive can use without or with minimum training• Operate by point- and click technology• Very rare use of command language

• time consuming and • difficult to learn and remember

• Must be customized for each executive

Using Standard Definition of Terms• Data dictionaries for transaction operations • Terms that are widely used should be standardize• To avoid misunderstanding/ misinterpretation

• sign-up by Lockheed• Dictionary of term used in the IES



Designing the Main Menu as a Gateway• As the “main door” to a variety of applications for executive

• e-mail, e-filing • external news, stock price

• To avoid unnecessary procedures, difficulty and inconvenience• Use of picks (menus, icons, buttons)

Designing the System for Ease of Navigation• Facilitate in moving from one to other application • Understanding the mental model of executive • The decision was business rather than technical• Navigation features:

• one screen shows where the user in the system• a home key or pick to back to the main menu• a single menu to provide direct access to different screens



Response Time as Fast as Possible• Monitor the response time of the system for each query• Trend to degrade (2.8 seconds in 1988 to 5.3 seconds in 1991)

• despite increase use of more powerful computer and LAN• satisfaction with response time was extremely low

• Response time will faster if the screens are static• Trade-off between timeliness of data on the screen

and response time• Executive usually tolerate a slow response to ad hoc queries

• still acceptable when compared to “conventional” method• to gather the same information




Part 3Document-Based Decision Support



• Huge amount of information generated and stored every second• Paper documents stored, need office floor space • Only 5 to 10 percent of paper based information is accessible• 40 to 60 percent of worker`s time entails working with documents• Document management to ease the access• Information of executives do not conform with traditional conception

of data queries• Information acquisition is centered on external and personal sources

non-numeric, text base, • A “ document” captures different sources of information:

numeric, non numeric, text, graphics, images, voice, video etc.• Document-based DSS (DDSS) expand the data component to gather,

store, manipulate, manage, and provide access to data• Traditional DSS failed in handling document-base data due technology

constraints• New technology(client-server network, telecommunication, storage, AI,

memory, open new possibilities in managing information

Introduction


The Data Subsystem

ExternalData bases

DSData Base

ExternalDocuments

InternalData bases

Internaldocuments

Data:• Extraction• capture• Entry• Links to other servers

DBMS Functions:• Creation-generation, restructure• Update data MBMS• Inquiry and retrieval• Update user profile


Technology Platforms

• DDSS inherently process large amount of data• Decreased cost and increasing processing speed• Technology supporting DDSS implementation:

• client-server technology• seamless access to distributed data and processing at any site• end user interact with network through graphical user interface

• storage technology• increasing capacity, new media (magnetic, optical, CD/RW)• integrate a variety of storage and processing technologies

• standard for input and process data• parallel processing

• speed-up access and processing time• relational data base,

• Oracle, Sybase, DB2, Informix


Technology Platforms

• online Analytical Processing (OLAP)• Pilot, Comshare, IRI• provide links to relational data base and object-oriented tools

• object oriented technology• Objectstore (object design), Gemstone (Servio)• Versant Object Technology• Objectivity

• imaging technology• scanning for picture and text• digitalization • enable to include any printed document in the DDSS• paperless environment

• Storage and input technology, powerful hardware• ease to use,• speed of access• flexible


Sofware Advances

Electonicmail

Onlinedatabases

Informational retrieval

Hypermedia

Document-BaseDecision Support

GroupwareDocument

ManagementSystem



Part 4Intelligent Support Systems



Introduction

Reduce product cycle time

Increase quality

Response to customer

Global competition

Decentralized smaller, flatter

improve communication reduce cost


Information Overload

Data volume & complexity

Time for processinganalysis, interpret

•To much data•Data may not be processed into useful information•No time to navigate all the data sources

Software agent/filter:•software robot•knowbots•cyberservants•intelligent assistant

Reduced data into meaningful information


Information Overload

Multidimensional

Hierarchical Perspectives:• products• locations• periods

Level of details:• products• locations• periods

The data

Process:•sort•calculate•compare•search

Soft data

• Amount of data collected doubles every year• Only 5 % can be analyzed knowledge worker

• 60 % try to discover the pattern of data• 20 % try to find the mean of the pattern• 10 % actually doing something about the pattern

• Information overload reduce decision making capability by 50%


Software AgentsThe Background

•Vanesser Bush`s vision (1945)• a machine enables user to navigate through the ocean of information

•John McCarthy (1950)• Soft robot, living and working in the computer network of information utilities

• Soft robot could perform appropriate action, ask forand receive advice from the user when it reach the dead -end


Software AgentsThe nature

Software agent

Operates in the background•user assign task, up to the agent to work• set and forget

Support conditional processing•rule-based, pattern-matching•the rule supplied by end user

Focuses on a single set of tasks•filtering e-mail•broken down the complex task

Automates repetitive task


Type of Agents

Batch Program• has a conditional logic and is run at scheduled interval• it possess the minimal set of attributes for an agent

Intelligent agent• have capacity to learn

• user habits• interest• behavior

• built with rule-base, script base language• only few available


Executive and Management Support System

Most of them non intelligent agent• Focus:

• monitoring, filtering, summarizing large amount or• frequently changing data set• automatically distributing the result to end user

• Architecture• a series of cooperative processes• agent is only one component

• Function• provide background intelligence• pass the matching pattern to the other process

• Intelligence is derived from a relatively small set of• expert system if/then/else/rules


Software Agent The definition

A background process that utilize a set of detection rules to automatically and routinely monitor the data set (hard/soft)for pattern defined by those rules to inform interested user when the pattern arise

software filtering agent / watcher agent

The data

0102030405060708090

Q1 Q2 Q3 Q4

Sales

The data

The data Detection Rules

Results


Architecture for Implementing Agent-Based ApplicationClient/Server Agent

Result

Request

Agent Server Client Application

• A client application• Data stored such away that they can be accessed by the

agent running on server• Agent is set up to evaluate the rule routinely• Return the results to the (same/other) client • Connection: use e-mail• Non mobile: the agent remains stationary on the server


Architecture for Implementing Agent-Based ApplicationRemote Programming

Result

Desktop

• Message (data & procedures) sent by one computer• Receiving computer executes the procedures

agent running on server• PDAs and AT&T personal link networks “agent go to place”• Operate under Unix: Tcl and Safe Tcl• Lack of support for remote programming• Virtually no implementation on IES

Desktop

Message


E-mail Filtering

• Artificial intelligent rules and graphical user interface to createintelligent e-mail assistant

• Automatically:• sort and store incoming message• alert• respond

• End user specify the request through keywords -phrases• Use IF - THEN rules• Can be found in almost every PC-based e-mail system


The Application of Agent Technology

Key of Success• Easy for user to specify rules• Easy to develop

• template, form• as macro, script or external program

• Potentially easier to create an intelligent e-mail filtering system• Maxims (MIT)• simplified learning process ( limited number of field, action)

• Information Executive System• To ease the problem of e-mail overload• To stay informed about the enterprises• To distribute the document to intended reader


Active Databases

Data

0102030405060708090

Q1 Q2 Q3 Q4

Sales

Data

Data

Decision Application

cleanedorganizedoptimized

Data warehouse

Data

Data


Data Base World (C/S)

Client• Graphical front-end for issuing queries to and receiving result • from data base server• Desktop

Server • Process or program that respond to the request• Host computer


Active - Passive Server

Classified by Stephen Schur• Whether the operation under external or internal control

Active Server • Has its own control logic• Operates on /is operated by:

• client• other server• system events

• Active Database “Robot for OLAP” (Comshore)

Passive Server• only perform operation requested by a client


Robot for OLAP

• Online analytical processing• Data stored in a multidimensional database system• Optimized for multi-dimensional comparisons• Robot for OLAP to search for trend and pattern• Report when matching

Applied by Hertz (Car rental)• Pricing analysis• Quickly respond to the competitor• Price leader: lowest price in a given location, market segment

and car class

Active Database agent• Filter numerical Data• Monitor the changes routinely


Competitive Intelligence from the Internet

• Using browser to view and transverse web pages • Direct access to online trade journal & newspaper• WWW -hypermedia document from data provider• Cheaper, timely, broad (often unique) access to• Different source of information

Competitive intelligence is based on learning, which is based on the ability to customers, consumers, to competitors,

to partners, to experts ….And most importantly to ones own enterpriseThe competitive environment send out the messages all the time: Tapping them and

learning from them is an art that requires open eyes, ears, mindsCompetitive environment send the message all the time

changes, trends, prospects, threats, weakness


Information from Internet

Distributed

Disorganized

Growing to fast

InternetArchieSpidersWorms Google

wanders

SearchEngine

Uncontrolled

Too much information

Hard to use


Concluding Remarks

Tremendous organizational and technological change• Explosion of the amount of data available for decision making• Intensified need for access the data at virtually any time, any place• Concomitant reduction in the amount of time and staff available for analyzing and understanding the data• Move toward team-base, globally distributed, decentralize entp.• Shortened product life cycle• Blurring the border between enterprise and their enviroment

• distributors, suppliers, partners• Increase mobility of knowledge workers• Computer for communication rather than for computation• Explosive rise of global computer network, internet


END