Application Abstracts

November 1st 1997

 CLIPS Application Abstracts
November 1st 1997

CONTENTS

Section 1 - Introduction ................................................................................................................1

Section 2 - Abstracts .....................................................................................................................3

Index.............................................................................................................................................47
 Section 1 - Introduction

This manual is the CLIPS Application Abstracts manual. This booklet contains brief descriptions,
supplied by CLIPS users, of applications where CLIPS is being used. Previous versions of this
manual were put together by Linda Martin, Wendy Taylor, Scott Meadows, and Ken Freeman.

If you’d like to share information with other CLIPS users about your CLIPS applications, please
provide the following information:

• The name of your expert system.

• Its purpose (brief—one or two sentences).
• Development stage (conceptual, developing, alpha testing, beta testing, finished).
• Other Languages/Shells used.
• Papers or other references that describe your application.
• Contact person (name, organization, address, phone, fax, email address, etc).
• Description of your application (no more than a page or so).

Send the information by electronic mail to [email protected].

Since the primary purpose of this booklet is to provide CLIPS users the opportunity to contact
other individuals developing applications of interest, we’re limiting the applications listed in this
booklet to those that provide at least some source of further information (such as a mail address,
email address, phone number, or reference other than the CLIPS conference proceedings). In
either case, we’re still interested in hearing about any applications you’ve developed.

1
 Section 2 - Abstracts

Expert System Name: Prototype Electronic Purchase Request (PEPR)

Purpose: To improve the efficiency of “paper” workflow systems by
automating commonly used forms in a commercial forms
package and having CLIPS validate the form and generate an
electronic “routing slip” based on its content.
Development Stage: Beta Testing
Other Languages/Shells Used: AppleScript, Informed Manager (commercial forms
package), FileMaker Pro (commercial DBMS),
PowerTalk/PowerShare (part of Apple’s Open Collaboration
Environment, aka “System 7 Pro”)
Last Update: April 28, 1994
Contact: Michael Compton,
AI Research Branch M/S 269-2
NASA Ames Research Center
Moffett Field, CA 94035-1000
Phone: (415) 604-6776
Fax: (415) 604-3594
Email: [email protected]

We have developed a knowledge-based system for improving the efficiency of automated

workflow systems by 1) ensuring the correctness and completeness of data contained on forms
that are originated and transmitted electronically, and 2) generating an electronic 'routing slip'
that reflects who must approve the form. The system uses a form-independent validation engine
and form-specific constraints to check that electronic forms are filled out correctly. If no errors
are detected during validation, the system uses information on the form to generate a list of
individuals and/or organizations that must approve it. This 'approval path' information is added
to the form so that it can be automatically routed from one approver to the next. The system is
implemented in CLIPS and currently runs on Macintosh computers. It communicates with an off-
the-shelf electronic forms package via AppleScript(tm) and can operate within the Apple Open
Collaboration Environment (AOCE™), which supports a variety of other workflow capabilities
including digital signatures, system-level electronic mail, and data encryption. The system has
successfully validated and generated approval paths for approximately ten different types of
forms, and is easily extended to new forms via a "BUILDCLASS" facility that automatically
generates the CLIPS code necessary to represent and reason about the new form.

3
Expert System Name: TENNIS
Purpose: To build an expert system to estimate and evaluate the ease
of service for a computer network.
Development Stage: Developing
Other Languages/Shells Used: C, XVT, DEC’s Polycenter NetView
Last Update: April 28, 1994
Contact: Dr. David C. Brown
Computer Science Department
Worcester Polytechnic Institute
Worcester, MA 01609
Phone: (508) 831-5618
Email: [email protected]

The system uses multiple phases, each consisting of a collection of small expert agents written in
CLIPS, to produce a set of service tasks and estimate their cost. The system connects to different
user interfaces for different types of users, and with different databases that hold descriptions of
hardware and software, costs, and a description of the network being evaluated.

Expert System Name: The Computer Aided Aircraft design Package (CAAP)
Purpose: engineeringTo aid in the preliminary design of modern, fixed
wing aircraft.
Development Stage: Developing
Other Languages/Shells Used: C and the Macintosh Toolbox
Last Update: April 28, 1994
Contact: Guy Yalif,
Princeton University
Department of Mechanical and Aerospace Engineering
114 Walker
Princeton, NJ 08544
Phone: (609) 258-7613
Email: [email protected]

CAAP is an expert system design to aid both the student and engineer design airplanes. Using a
custom standard “Macintosh” user interface “look and feel,” CAAP allows the user, based on
simple specifications, to perform preliminary airplane design. When presented with results, the
user can then change the configuration and see the effect their change has on the airplane. For
example, if CAAP designs a plane with a 40 foot wingspan and you realize that your hangar is
only 30 feet wide, you can shorten the wing and see the effects this change has on the rest of the
plane.

4
Expert System Name: Intelligent Remote Automation Project
Purpose: Embedded automation in PLC / C Modules for field
automation.
Development Stage: Conceptual / Developing
Other Languages/Shells Used: C++, C, G2, WindExS, Level5
Last Update: April 28, 1994
Contact: Edward B. Toupin,
Texaco Trading & Transportation, Inc.
Email: [email protected]

The application will reside in remote racks made part of a wide area SCADA system. The expert
system will provide intelligence at a low level in the system's hierarchy for immediate decision
making as well as automated control in case of loss of communications with central control.

Expert System Name: PROMEAT

Purpose: Quality inspection in food processing industries.
Development Stage: Developing
Other Languages/Shells Used: ET++
Last Update: April 28, 1994
Contact: Thomas Cord
Forschungszentrum Informatik (FZI)
Technical Expert Systems and Robotics
Haid-und-Neu-Strasse 10-14
76131 Karlsruhe, Germany
Phone: ++49/721/9654-322
Fax: ++49/721/9654-309
Email: [email protected]

Cured ham of Parma is well known to be a delicacy. The unique taste of this ham is guaranteed
by the high quality of raw material, the scrupulousness and experience during production as well
as the exceptional climatic conditions in the region of Parma.

The “Consorzio Carni Suine Garantite” from Reggio Emilia in Italy and the spanish company
“Oscar Mayer Alimentacion” in Valencia will apply a quality control system, that is developed in
the european BRITE/EURAM project PROMEAT, in their ham production in order to increase
the quality of the final products and the efficiency and economy of the meat manufacturing
process. Objectives of the system are the supervision of the different production phases and the
optimization of product quality.

Common to all processes in the meat transformation industry are the biochemical complexity and
the high cost of the raw material. It is very difficult to make objective measurements of quality

5
parameters through non-destructive techniques. This all supports the view that human experts are
central to supervision of production and therefore cannot be replaced by machines.

A knowledge-based system to support process supervision and quality control operations is

developed. The data acquired by different sensors, the parameters and features extracted in-
process, and additional observations supplied by the operator are collected and combined. The
interpretation of these data allows the determination of the process state and the recognition and
diagnosis of product defects. The process supervision system is based on a blackboard
architecture, whose knowledge base represents the properties of the manufactured products, the
production phases and the devices involved in production.

Expert System Name: TOPKAT (The Open Practical Knowledge Acquisition

Toolkit)
Purpose: To build a tool which supports structured knowledge
acquisition techniques (transcript analysis, repertory grid,
card sort, laddered grid) and knowledge modelling using the
CommonKADS method.
Development Stage: Prototype finished. Final system under development.
Other Languages/Shells Used: HARDY (Hypertext and Diagramming System)
Last Update: April 28, 1994
Contact: John Kingston,
AIAI
Edinburgh
Email: [email protected]

TOPKAT (The Open Practical Knowledge Acquisition Toolkit) is a hypertext and diagram-based
toolkit which integrates knowledge elicitation techniques with the CommonKADS approach to
knowledge modelling.

The KADS methodology for developing knowledge-based systems attempted to resolve this
problem by suggesting that knowledge should be analysed on several different levels
simultaneously: the domain level, the inference level and the task level. The development of a set
of generic models which serve as a template for the inference level of knowledge has proved
extremely useful, with the result that KADS is now the most widely used methodology for KBS
development in Europe. CommonKADS, the recent successor to KADS, has extended and
refined the recommended representations for each level, so that CommonKADS now provides a
comprehensive suite of representations for the analysis of knowledge. In particular,
CommonKADS has defined a set of ontological primitives with which domain level knowledge
can be analysed.

CommonKADS aims to provide a detailed and re-usable approach to the analysis of acquired
knowledge. It is therefore important to understand how the results of various knowledge
elicitation techniques should be mapped onto CommonKADS models; however, there is

6
currently little understanding of how the results of structured approaches to knowledge
elicitation, such as card sorting or the repertory grid, could be mapped to other formalisms.
Integration between structured knowledge elicitation techniques and the CommonKADS
modelling methods is being developed using a hypertext and diagram-based toolkit, known as
TOPKAT (The Open Practical Knowledge Acquisition Toolkit). TOPKAT includes support for
knowledge elicitation techniques (transcript analysis, laddered grids, card sorting and repertory
grids), support for developing the CommonKADS model of expertise, and hyperlinks between
different representations. A particularly useful feature is the ability to translate from
CommonKADS to the knowledge elicitation tools; this allows one knowledge elicitation
technique to generate input for another.

TOPKAT is implemented in HARDY and in CLIPS; a number of functions have been written
which allow CLIPS to create, modify, access or delete HARDY diagrams and hyperlinks. This
allows CLIPS to be used for much of the analysis of the acquired knowledge, such as the
statistical comparison of elements in the repertory grid. HARDY (and therefore TOPKAT) runs
on machines supporting either X Windows or Microsoft Windows.

Expert System Name: COURSE SELECTOR

Purpose: To allow students to choose courses which comply with
university regulations.
Development Stage: Finished
Other Languages/Shells Used: None
Last Update: April 28, 1994
Contact: John Kingston,
AIAI
Edinburgh
Email: [email protected]

The COURSE SELECTOR system was implemented for the Department of Business Studies in
the University of Edinburgh. The Department's problem was that, in the first two weeks of the
Autumn term, every student is required to choose courses for the coming year. Each student has a
Director of Studies who is responsible for ensuring that a legitimate combination of courses has
been chosen, and every Director of Studies finds that the whole of the first week of term, plus a
significant proportion of time thereafter, is taken up with advising students on this complex
problem. The task of choosing an acceptable combination of courses is complex.

The current procedure (in theory) is for the students to examine the University Calendar, an 800-
page volume describing the regulations and timetables of every available course, and to make
their course choices which are then verified by their Director of Studies. In practice, many
students rely on their Director of Studies to be a source of wisdom, making little or no effort to
look at the University Calendar themselves. The result is that the Director has to conduct one or
more lengthy interviews with each student. Since each Director is currently responsible for 60
students, the workload is large. There is also considerable scope for error; the number of possible

7
interactions between courses is so great that, during the development of the course selector
system, the University Calendar itself was found to have omitted to mention a timetable clash
between two courses which were recommended for a particular degree. The COURSE
SELECTOR system was designed to encode the knowledge stored in the University Calendar,
with some additional input from two experienced Directors of Studies.

A version of the KADS methodology was used to build the system; the final stages of this
method recommended that the system use data driven reasoning, object oriented representation,
truth maintenance, and external file storage. A brief description of the actual design of the system
is given below:

• Data driven reasoning was implemented using forward chaining rules.

• The recommended object-oriented representation was actually implemented using facts,

where each object was represented by several facts. The first element in each fact was the
name of an “object.”

• Truth maintenance was implemented using facts, since the built-in truth maintenance facility
was not sufficiently expressive. Truth maintenance was implemented by the simple but
powerful technique of creating a fact to represent a course which was known NOT to be
eligible for selection. This contrasts with the normal truth maintenance technique of keeping
track of valid assumptions; the reason for this choice was that there are likely to be fewer
ineligible courses than eligible ones, and so fewer “truth maintenance” facts will be
required, leading to increased efficiency. The “truth maintenance” facts note the reason for
the creation of the fact, which will be the addition of a certain course to the course schedule;
if that course is ever removed from the schedule, then any “truth maintenance” facts
associated with it are also removed. This technique is powerful because it is able to handle a
situation where a course is ineligible for more than one reason; a course is only considered
eligible if all the “truth maintenance” facts affecting it are removed.

• The external file of course information was developed by using a spreadsheet, and writing
out a text file containing the fields of the spreadsheet. This file was then parsed using an
ASCII parser.

In addition, it was decided that the KBS should be broken down into separate files of rules, and
that the content of these files should mirror the functional decomposition (and hence the models
of expertise and interaction) as far as possible. This decision helped in the debugging of the
KBS, and clarified later decisions about where to store certain rules.

8
Expert System Name: GermWatcher
Purpose: medicalTo assist the Infection Control Department of Barnes
Hospital (a large teaching hospital affiliated with the
university) with its infection control activities. These
activities include surveillance of microbiology culture data.
Development Stage: Used in production since February 1993
Other Languages/Shells Used: Sybase ISQL scripts, Bourne shell scripts
Last Update: April 28, 1994
Contact: Sherry Steib,
Washington University School of Medicine
Division of Medical Informatics
660 South Euclid
Campus Box 8005
St. Louis, Missouri 63110
Phone: (314) 362-4322
Email: [email protected]

Hospital-acquired (nosocomial) infections represent a significant cause of prolonged inpatient

days and additional hospital charges. We have developed an expert system called GermWatcher,
which applies the Centers for Disease Control's (CDC) National Nosocomial Infection
Surveillance (NNIS) culture-based criteria for detecting nosocomial infections. GermWatcher
has been deployed at Barnes Hospital, a large tertiary-care teaching hospital, since February
1993.

Microbiology culture data from the hospital's laboratory system are monitored by GermWatcher.
Using a rulebase consisting of a combination of the NNIS criteria and local hospital infection
control policy, GermWatcher scans the culture data, identifying which cultures represent
nosocomial infections. These infections are then reported to the CDC.

9
Expert System Name: GermAlert
Purpose: medicalTo assist the Infection Control Department of Barnes
Hospital (a large teaching hospital affiliated with the
university) with its infection control activities. These
activities include surveillance of microbiology culture data.
Development Stage: Used in production since February 1993
Other Languages/Shells Used: Sybase ISQL scripts, Bourne shell scripts
Last Update: April 28, 1994
Contact: Sherry Steib,
Washington University School of Medicine
Division of Medical Informatics
660 South Euclid
Campus Box 8005
St. Louis, Missouri 63110
Phone: (314) 362-4322
Email: [email protected]

Most hospitals have infection control programs which are aimed at the early detection and
aggressive treatment of infections. The earlier an infection is discovered and treated, the less
likely it is to spread to other patients and hospital staff--and the less likely it is to prolong the
infected patient's stay in the hospital. We have developed an expert system called GermAlert,
which applies local hospital culture-based criteria for detecting “significant” infections, which
require immediate treatment. GermAlert has been deployed at Barnes Hospital, a large
tertiary-care teaching hospital, since February 1993.

Microbiology culture data from the hospital’s laboratory system are monitored by GermAlert.
Using a rulebase consisting of criteria developed by local infectious diseases experts, GermAlert
scans the culture data and generates an “alert” to the Infection Control staff when a culture
representing a “significant” infection is detected.

10
Expert System Name: Drive Analysis Expert System
Purpose: Scan log files from diagnostics and OS after running tests on
hard drives and determine their pass/fail status.
Development Stage: Finished
Other Languages/Shells Used: C, C-shell
Last Update: April 28, 1994
Contact: Greg Moore
Mack Technologies, Inc.
27 Carlisle Rd.
Westford, MA. 01886
Phone: (508)-392-5539 (days)
Email: [email protected]

In a high volume drive test (Winchester hard drives) process, we can test anywhere from 1 to 20
drives. When testing is completed, the expert system reads the log files generated by the
diagnostics, the operating system and a serial number collection routine.

The system generates 3 reports:

1) PASS/FAIL status: Each drive is identified (serial number) and tagged with a pass or fail
status. If the drive failed, a general failure description is given. This is a single page report
with the target audience being a technician who may or may not have any experience with
the particular test environment and/or Winchester technology.

2) Summary of generated errors: This is a summary of all error activity for each drive tested.
The report gives more detail than the pass/fail report above, and is targeted to an engineering
level audience. In our application, this page is usually given to the drive manufacturer along
with the drive for failure analysis.

3) Unknown situation report: This is a report used by the developer of the expert system (me)
to alert when a situation has occurred that the expert system could not handle. Corrective
action can then be taken to handle the problem the next time it occurs.

In addition to the above three reports, we save the log files for future reference in resolving
issues from report #3 above and as a basis to test future changes to the rule base.

The system has helped by:

• Reducing the level of knowledge required by the person operating the test. Allows higher
level personell to be utilized elsewhere, while allowing a greater pool of personell to attend
the test process

11
• Eliminating the risk of shipping a bad product. A bad drive can generate megabytes of data.
If a failure on another drive is embedded within this massive data, chances are a person
would miss the entry.

• Reduced the time required to disposition drives after test.

• Alerted engineering to trends sooner than with the manual process.

• Saves paper - most final report runs are 3 pages verses the hundreds of pages that the log
files may take up.

The system was developed in Unix, on a Sun platform running SunOS 4.1.3. We have since
turned the process over to Solaris 2.3.

Expert System Name: FIRE-XPS

Purpose: Analysis and Diagnosis of the setup of complex fluid flow
calculations using the commercial CFD code FIRE (TM of
AVL)
Development Stage: Developing
Other Languages/Shells Used: FORTRAN, C
Last Update: April 28, 1994
Contact: Peter Blahowsky,
AVL List GmbH
Fluid Dynamics Research
Kleiststrasse 48
A-8020 Graz
Austria
Phone: xx43 316 987-618
Fax: xx43 316 987-777
Email: [email protected]

The setup of a modern CFD calculation is a quite formidable task and a successful flow
calculation depends heavily on the users experience. Especially for new users or new
applications it is necessary to support the user of FIRE with in depth knowledge about fluid flow
problems in general, numerical, calculation mesh quality and stability criterias. The choice of
the right parameters like time step or under-relaxation factors determines the speed of a flow
calculation. FIRE-XPS will support the user of taking the right choices of parameters and
boundary conditions, check the calculation mesh quality and analyse flow calculations which did
not succeed at all. The knowledge and experience of the development and application group will
thus be easier accessible to the user. As a side effect we expect to reduce training time for new
users. FIRE-XPS will be a stand alone process exchanging information with FIRE via sockets or
RPC.

12
Expert System Name: ITS-Engineering
Purpose: The development of an Intelligent Tutoring System Shell for
Engineering. The system contains pedagogical teaching
styles for engineering, student models of engineering
learning, and a knowledge representation for technical
subjects.
Development Stage: Conceptual / Developing
Other Languages/Shells Used: C, TAE+, multimedia interfaces
Last Update: April 28, 1994
Contact: Nelson Baker,
Assistant Professor
Georgia Institute of Technology
School of Civil and Environmental Engineering
790 Atlantic Drive
Atlanta, Georgia 30332-0355 USA
Phone: (404) 894-2223
Fax: (404) 894-2278
Email: [email protected]
Mosaic: http://howe.ce.gatech.edu/Faculty/Structures/
N.Baker/baker.html

Researching, developing, and deploying new and creative methods for disseminating and
educating diverse populations of engineers and scientists (practitioners as well as students)
comprises the motivation for creating ITS-Engineering. Continued life long education is a
reality of today's engineering work place and methods are needed to provide this education; thus,
the motivation for doing the research necessary to create this system.

Efforts are directed to determine how techniques of artificial intelligence, multimedia, computer
simulation and visualization, and engineering pedagogy can be combined or enhanced to
stimulate students, help them to excel in their studies, while retaining and improving quality
education.

Through these activities, research and development are underway to investigate the learning
styles of engineers along with pedagogical teaching techniques for engineering material. What
kinds of experiences benefit engineering instruction for some people? Can these experiences be
provided via software to increase the learning comprehension of others? Assuming that software
systems will benefit engineering education, the time to develop the software is very lengthy.
Thus, ITS-Engineering is looking into ways to produce a shell which incorporates sound
pedagogical strategies and student learning methodologies. The shell can then be used to
develop applications which require only the encoding of the subject domain.
Research/development is approaching these issues by studying how to combine multimedia with
knowledge-based systems; how to automatically generate examples and exercises for the student

13
to achieve pedagogical objectives; how to incorporate appropriate levels and timing of
explanation to users; how to allow the spatial visualizations needed by most engineers during
problem solving, namely sketching; how to instruct engineers using this new technology; and
how to deploy these systems when completed.

Expert System Name: Event Pattern Language (EPL)

Purpose: Extend Sybase trigger mechanism with temporal information
Development Stage: Beta Testing
Other Languages/Shells Used: Sybase
Last Update: April 28, 1994
Contact: Giovanni Giuffrida
UCLA
11050 Strathmore Dr. #403
Los Angeles, CA, 90024
Phone: (310) 285-2476
Email: [email protected]

Description of your Application: EPL allows the user to specify rules modelling temporal
sequences of generic events (insert, delete or update) on a database. Actions can take place when
a certain pattern of events is completely satisfied. The basic idea is that you can model situations
like:

1) “If the temperature goes down for 3 consecutive days then ...”
2) “If there is a withdrawal for more than $100,000 and within the next 20 minutes a deposit
into the bank account for the same amount of money then ...”
3) “If IBM share goes down for the entire week then ...”

Each of the previous examples involve sort of temporal relationships (enriched with constrains
on some arguments) on database accesses. The current implementation uses SYBASE as
database host. EPL itself is built on top of CLIPS which is in turn coupled with SYBASE
through Open-Library.

14
Expert System Name: Numerical Propulsion System Simulation (NPSS) Monitor
Purpose: To monitor a variety of engine component simulations and
the data exchange across various computing platforms.
Development Stage: Conceptual
Other Languages/Shells Used: None
Last Update: April 28, 1994
Contact: Henry Lewandowski
NASA Lewis Research Center
Cleveland State University
Industrial Engineering Dept.
Cleveland, Oh 44115
NASA Phone: (216) 433-6542
CSU Phone: (216) 687-4668
Email: [email protected]

The NPSS project is a joint effort between NASA, university researchers and industry to bring
advanced design analysis techniques to the next generation of propulsion systems. As a part of
this effort, an expert system is being designed to monitor the various engine component
simulations as they run and to examine the data that is exchanged between modules. The various
models run on heterogeneous platforms in a parallel environment.

Expert System Name: PRISM

Purpose: Price quotation system for telephone switches.
Development Stage: Alpha Testing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: William Mettrey,
Bell Northern Research
P.O. Box 13478
Research Triangle Park, NC 27709-3478

We are testing a Price Quotation System that has approximately 500 rules and is expected to
grow significantly.

15
Expert System Name: None
Purpose: An expert diagnostic and maintenance system.
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Albert Koval
Cray Research
P.O. Box 17500
Colorado Springs, CO 80935

CLIPS is the cornerstone for an expert diagnostic and maintenance system being developed for
the Cray III.

Expert System Name: None

Purpose: Planning and scheduling and intelligent databases.
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Steven Gold
General Electric
780 Third Ave.
Bldg. 8, Room 8347
King of Prussia, PA 19406

We are using CLIPS in Planning and Scheduling. Also, we are using CLIPS in intelligent data
bases. We are running on a Sun III with OS 4.

Expert System Name: None.

Purpose: Helps people to use SLATEC routines.
Development Stage: Beta Testing
Other Languages/Shells Used: SLATEC
Last Update: May 29, 1992
Contact: Barbara Helland,
Iowa State University
Ames Laboratory
136 E. Wilhelm
Ames, IA 50011

We automated a document search using the SLATEC mathematical subroutine package, and
added an Expert System that will help people to use SLATEC routines.

16
Expert System Name: None
Purpose: A consultant for chemical engineers who design chemical
plants or refineries.
Development Stage: Developing
Other Languages/Shells Used: ASPEN
Last Update: May 29, 1992
Contact: W.J. Parkinson, W.J.
Los Alamos National Lab
MS: G787
P.O. Box 1663
Los Alamos, NM 87545

We are developing an Expert System to be used with ASPEN - a chemical process plant flow
sheet simulator. It will be a consultant for young chemical engineers who design chemical plants
or refineries.

Expert System Name: Laser Docking Sensor Associate

Purpose: Assists mission specialists aboard the Space Shuttle in
rendezvous/docking and space experiments.
Development Stage: Finished
Other Languages/Shells Used: TSR Graphics
Last Update: May 29, 1992
Contact: Paul Griffith,
Microexpert Systems, Inc.
24007 Ventura Blvd. #210
Calabasas, CA 91302
Phone: (818) 712-9934

We have developed the Laser Docking Sensor Associate. It has a TSR Graphics Interface with
CLIPS. There are approximately 80 compiled rules for checking data evaluation, anomaly
checking and vision analysis. It runs on a Grid Shuttle Payload general support computer.

Developed for NASA Johnson Space Center, the Laser Docking Sensor Associate assists mission
specialists aboard the Space Shuttle in rendezvous / docking and space experiments. It integrates
three Expert Systems for multiple, simultaneous, real-time graphical visualizations of data from
laser docking sensors to aid in maneuvering the spacecraft.

17
Expert System Name: Planners Workbench
Purpose: Integrates knowledge gained from shop floor experience
with computerized CAD/CAM data to support cable
assembly planning.
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Paul Griffith,
Microexpert Systems, Inc.
24007 Ventura Blvd. #210
Calabasas, CA 91302
Phone: (818) 712-9934

Planner’s Workbench integrates knowledge gained from shop floor experience with
computerized CAD/CAM data to support cable assembly planning. It incorporates manufacturing
expertise into the design, layout, color-coding, hardware placement, and assembly of cables for
the Navy's Trident missle systems. It runs on an IBM RT workstation.

The Expert System modules are as follows:

• Data Convertor Module

• Reads dimensions from CADAM file.
• Recognizes cable components in CADAM file.
• Reads cable geometry.
• Links CADAM data with wires and parts data to create the database.

• Form Board Design Module

• Arranges harness on board.
• Eliminates crossovers and bends to fit.
• Considers human and machine reach and access, bending radii and supports.
• Allows user to override expert system bends with custom bends.

• Connector/Color Coding Planning Modules

• Sequences connector assembly.
• Minimizes wire ambiguity.
• Optimizes movement of continuity/Hi Pot tester equiptment.
• Assigns color coding to individual wires.

• Wirematic Planning Module

• Fits cable in minimum space.
• Identifies guide and anchor pin locations.
• Creates full-scale CADAM drawing.

18
 • Creates wirematic routing program.
• allows user to override Expert System bends with custom bends.

• Hardware Assembly Planning Module

• Sequences cable segment assembly.
• Organizes cable segments into wire bundles.
• Minimizes loose wire bundles.

• Report Generating Module

• Generates assembly method instructions for operators.
• Generates programs for wirematic and cablescan support equiptment.
• Makes full-scale CADAM drawings for form board and layup board.

Expert System Name: STRUTEX

Purpose: Initially configures a structure to support point loads in two
dimensions.
Development Stage: Finished
Other Languages/Shells Used: FORTRAN, DI-3000, RIM, PROLOG
Last Update: May 29, 1992
Contact: James Rogers,
NASA Langley Research Center
MS: 246
Hampton, VA 23665-5225

We have developed a prototype knowledge base system for initially configuring a structure to
support point loads in two dimensions. This system combines numerical and symbolic
processing by the computer with interactive problem solving aided by the vision of the user by
integrating a knowledge base interface and inference engine, a database interface, and graphics -
while keeping the knowledge base and the database files separate. The system writes a file
which can be input into a structural synthesis system, that combines structural analysis and
optimization.

STRUTEX emulates an engineering student taking a blank sheet of paper to a teacher to discuss
an idea for building a structural model to support one or more point loads in two dimensions. As
the teacher asks questions about the loading conditions and the support surface, the student
responds with answers or by sketching ideas on the piece of paper. Based on what is seen and
heard the teacher can help the student determine a reasonable initial structure for supporting the
given loads. In STRUTEX, a knowledge base replaces the teacher, a graphics window on the
computer replaces the piece of paper, and a dialogue area in the graphics window replaces the
verbal question and answer. The user interactively interfaces with the system through two
methods, typed dialogue and mouse-oriented graphics. The user graphically inputs loading and
support surface data using the mouse in response to questions about the load points, support
surface, and support structure. The data is stored in a relational database.

19
Once all questions are answered, the appropriate data is transferred from the database to the
knowledge base and the system determines the type of structure most suitable for satisfying the
input conditions. If the structure is determined to be a beam(s) or a string(s), then the structure is
drawn on the graphics window and the session is completed. If there is only one load and the
structure is determined to be a truss, then other rules are invoked to determine whether or not
bracing is needed, and, if so, the type and amount of bracing. This structure is then drawn on the
graphics window. If there is more than one load point and the structure is determined to be a
truss, then the user is guided by recommendations in a step-by-step approach to building the
truss. The truss built by the user is then tested against rules in the knowledge base and
recommendations are given for the user to improve the model. This is done iteratively until all
rules are satisfied and no recommendations for improvements are made. An input file for a
structural analysis program is written for a truss so that the model can be analyzed and optimized
by a previously developed system for structural synthesis.

One objective was to investigate methods for passing data between a database and a knowledge
base. This was accomplished by separately integrating two types of inference engines, one
forward chaining based on production rules, and one backward chaining based on PROLOG, into
the system and determining their effects on the flow of data between the knowledge base and the
database. No significant problems were encountered in integrating either of the inference
engines. Nor did one inference engine run significantly faster than the other for this small
knowledge base. It was concluded that these two systems supplement rather than compete with
one another.

A second objective was to examine when it is preferable for a computer to supply the data and
when it is preferable for the data to be supplied by human vision. It was also concluded during
the development of this system, that there are times to rely on the computer and there are times to
rely on the vision of the user. For small problems such as the ones used for testing, there are
several instances where the user's vision was more preferable than relying on the computer, such
as determining the location of the support surface relative to the loads. However, for larger,
more complex problems, it might be preferable to add symbolic rules to the knowledge base,
numerical algorithms to the main program, and rely on the computer.

The main program driver for STRUTEX is written entirely in Fortran. Other components were
added by linking existing software - DI-3000 for the graphics, RIM (relational Information
Management) for the relational database management, and CLIPS for the inference engine - to
the main driver program. The data for RIM and the knowledge base (rules) for CLIPS are
maintained in different files separated from STRUTEX. EAL (Engineering Analysis Language)
for the structural analysis, and CONMIN (Constraint Minimization) for the optimization are
coupled in PROSSS (Programming System for Structural Synthesis) to perform the analysis and
optimization.

20
Expert System Name: None
Purpose: Implements a decomposition scheme suitable for multilevel
optimization and to display the data in an N x N matrix
format.
Development Stage: Finished
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: James Rogers,
NASA Langley Research Center
MS: 246
Hampton, VA 23665-5225

CLIPS was used to develop an engineering knowledge based tool for decomposing complex
design problems into a suitable multilevel structure based on the multilevel optimization
approach. This tool requires an investment of time to generate and refine the input for each
design module. This investment may not be justified for a small, well-understood problem, but
should save a significant amount of money and time in organizing a new design problem where
the ordering of the modules is still unknown. The decomposition of a complex design system
into subsystems requires an interaction with the judgment of the design manager. This tool can
aid the design manager in making decomposition decisions early in the design cycle.

This tool provides help to the design manager by reordering and grouping the modules based on
the links (interactions) among the modules. The modules are grouped into circuits (the
subsystems) and displayed in an N x N matrix format. The feedback links, which indicate an
iterative process, are limited and restricted to be within a circuit. Since there are no feedback
links among the circuits, the circuits can be displayed in a multilevel format. Thus, a large
amount of information is reduced to one or two displays. The displays are stored and can be
easily retrieved and modified. The design manager and leaders of the design teams are given a
visual display of the design problem and the intricate interactions among the different modules
so that they can see how a change in one subsystem will effect other subsystems. It also helps
reduce the possibility of overlooking important links.

The tool gives the design manager the capability of examining the potential savings in time by
executing some of the modules in a circuit in parallel. A substantial time savings can be
obtained if circuits on the same level of the multilevel structure are executed in parallel. The
time savings as well as the number of processors that will be required are determined. In
addition to decomposing the system into subsystems, the tool examines the dependencies of the
problem and creates a dependency matrix. This matrix shows the relationship among the
independent design variables and the dependent objective and constraint functions.

21
Expert System Name: BRAT
Purpose: Simulation of the actions of the man-in-the-loop in
conducting command and control tasks.
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Brett Gossage
MS: 1-1-188
Nichols Research Corporation
4040 S. Memorial Pkwy.
Huntsville, AL 35802

We are building BRAT - BMC3 Requirements Analysis Tool - a simulation tool for the U.S.
Army Strategic Defense. It will simulate actions of the man-in-the-loop in conducting command
and control tasks. Development is on a Compaq 386, with Dos and Microsoft C 5.1, and on a
Vax VMS with Dec C.

Expert System Name: None

Purpose: An Expert System to manage the operation of the Space
Shuttle’s fuel cell cryogenic reactant tanks.
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Amy Murphey
Rockwell Space Operations Co.
MC: R16G
600 Gemini Ave.
Houston, TX 77058

To ensure an adequate and uninterrupted supply of electrical power during all phases of the
Space Shuttle's flight, it is necessary to judiciously monitor and control the flow of cryogenic
hydrogen and oxygen out of storage tanks and into the fuel cells. To maintain a reliable supply of
cryogenic reactants and to provide redundancy and fault-tolerance, the Power Reactant Storage
and Distribution (PRSD) tanks must be depleted as evenly as possible, within the constraints of
certain guidelines. Therefore, the PRSD system must be configured at certain times during the
nominal mission according to, not only the mission profile and initialization specifications, but
also an appropriate schedule of tank depletion.

We are developing a rule-based expert system which may be used for flight design to manage the
operation of the Space Shuttle's PRSD system. The expert system provides the user with
recommendations on how to configure the PRSD system. That is, for a given state of the PRSD
system, the expert system indicates which manifold valves to close and which tanks to activate.

22
The knowledge of this expert system is based upon standardized management criteria established
by Johnson Space Center. Thus, this expert system is a tool to aid flight design analysts of the
Space Shuttle's Electrical Power System to devise PRSD operational schemes during preflight
planning.

Expert System Name: EMI Detection Expert

Purpose: An EMI detection Expert System.
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Dave Swanson,
SEI
4241 Jutland
San Diego, CA 92117

We are developing an EMI Detection Expert System for the U.S. Navy.

Expert System Name: Woodpecker

Purpose: Assists in the management of woodpecker habitat.
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Robert Coulson,
Department of Entomology
Texas A&M University
College Station, TX 77843

No abstract.

Expert System Name: Weed Control Advisor

Purpose: Selects proper herbicides for weed control in rice production.
Development Stage: Finished
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Robert Coulson,
Department of Entomology
Texas A&M University
College Station, TX 77843

The Weed Control Advisor provides expert advice on the selection of effective herbicides during
all phases of the rice production time-line. It also provides the following:

23
• Explanations of why a particular herbicide is or is not suggested.
• Effectiveness of appropriate herbicides.
• Application information such as rates of applications and general application methods.
• Information on how a suggested herbicide provides control against weeds.
• Any warnings that might apply to a particular herbicide.

It is also a learning tool and can be used by those who want to learn more about weeds common
to rice fields, and their control. Users can learn more about herbicides, when they are most
effectively used, rates of use, and other application information by using the Weed Control
Advisor.

Expert System Name: Southern Pine Beetle

Purpose: Forecasts infestation of southern pine beetles.
Development Stage: Developing
Other Languages/Shells Used: GIS, GRASS
Last Update: May 29, 1992
Contact: Robert Coulson,
Department of Entomology
Texas A&M University
College Station, TX 77843

Forecasts infestation of southern pine beetles based on information about the area considered,
such as the age of the forest, the occurrence of lightning, and the history of bark beetles.

Expert System Name: INSEX

Purpose: Recommends insecticides to use in forest management.
Development Stage: Beta Testing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Robert Coulson,
Department of Entomology
Texas A&M University
College Station, TX 77843

No abstract.

24
Expert System Name: ISPPEX
Purpose: Integrated pest management system for southern pine
beetles.
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Robert Coulson,
Department of Entomology
Texas A&M University
College Station, TX 77843

ISPBEX, Integrated Southern Pine Beetle Expert System, is a software program which uses
Artificial Intelligence programming techniques to solve problems associated with identifying and
suggesting treatment recommendations for the control of southern pine beetle infestations. The
system utilizes information about specific pine beetle spots which can be entered directly through
the data entry screens in ISPBEX and transferred to or retrieved from the Data General. The
Expert System asks you, the user, questions about your particular problem and utilizes field data
collected from spots during ground checking. Rules based on knowledge compiled from experts
specializing in forest management, wildlife management, and southern pine beetle biology are
applied to the field data to determine the appropriate treatment actions. ISPBEX also provides:

• Data entry and update capabilities for SPBIS and RCW data sets.
• Spot growth model for estimating tree mortality up to 31 days.
• Spot geometry routine for estimating the distance the spot will travel up to 1600 meters.
• Treatment priority classification for the spot.
• Explanations of why a particular management tactic was recommended.

ISPBEX is also a learning tool and can be used by those who want to learn more about SPB
(Southern Pine Beetle) management. It is a menu driven system which queries the user for input
or selection of menu items. There are two ways to respond to these queries. The first way is to
select items from a list of choices that ISPBEX presents to you. The second way is to type
values into menus that resemble SPBIS and RCW data forms.

25
Expert System Name: UNIX Advisor
Purpose: Assist Unix Microcomputer System Administrators in
diagnosing and solving problems.
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Mott Given
U.S. Army
MS: DSAC-TMP
Defense Logistics Agency
Bldg. 27, Section 1, P.O. Box 1605
Columbus, OH 43216-5002

I am developing UNIX Advisor: an Expert System to assist UNIX system administrators for
Gould 9050 minicomputers running UTX/32 UNIX (which is based upon BSD 4.3). The system
covers the following types of problems:

• RJE problems, eg. line down, lines up but data not being transmitted.
• System crashes, hangs, panics, or halts.
• Tuning advisor.
• Analyzing system console messages.
• Why is the file system filling up?
• How do you rebuild a crashed disk?
• Building a facility for people to keep online notes about problems.
• Sources of information on UNIX.
• Automatically run certain monitoring commands.

In its present form, UNIX Advisor covers the initial requirements, as well as having the
following:

• A facility to let users exit to the UNIX shell, run UNIX commands, and return to the Expert
System.
• A survey form that application users go through as they exit the application.
• A note-taking facility for recording observations about how different types of problems that
are not currently in the knowledge base were solved.
• An electronic mail facility to send comments to the author of UNIX Advisor.

26
Expert System Name: None
Purpose: Assists in management decisions for steep pastures in the
Appalachian region.
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Douglas Boyer
USDA-ARS
P.O. Box 867
Beckley, WV 25802-0867

CLIPS is being used to build an Expert System for assisting in management decisions for steep
pastures in the Appalachian region. Given certain soils, macro-climate, micro-climate,
management goals and farmer experience, CLIPS will assist the farmer in making decisions that
will make the best use of his land, with resource conservation being a requirement.

Expert System Name: Front-end to GIS

Purpose: Acts as an aid and enhancer of the GIS decision making
process.
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Robert MacArthur,
Computer Application Group
College of Agriculture
University of Arizona
Tucson, AZ 85721

No abstract.

Expert System Name: None

Purpose: Verification and validation of expert systems.
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Peter Green
Worchester Polytechnic Institute
100 Institute Road
Worchester, MA 01609

No abstract.

27
Expert System Name: ASW Info Sys Dynamic Data Store
Purpose: Test & track file storage
Development Stage: Conceptual
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Dr. John Esch,
Paramax Systems Inc.
P.O. Box 64525
MS - U1N28
St. Paul, MN 55164

Prototype / Test both dynamic track file storage in a distributed ASW Inf. Sys and CLIPS.

Expert System Name: SelectPC

Purpose: Allows selection of PC equipment on the basis of a client’s
software specifications.
Development Stage: Conceptual
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Rob Schaller
Lockheed
2400 Nasa Road One
MS - B08
Houston, TX 77058

No abstract.

Expert System Name: Leukemia Diagnostic Assistant

Purpose: Assist clinicians in analysis of data from cancer patients.
Development Stage: Conceptual
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Gary Salzman
Los Alamos National Lab
MS - M888
Los Alamos, NM 87545

Flow cytometry data is obtained from the patient. System then supervises cluster analysis and
applies rules to make decisions about the various cell populations.

28
Expert System Name: None
Purpose: Medical expert systems
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: James M. Lamiell
Brooke Army Medical Center
Box 4, Bldg 1029
Fort Sam Houston, TX 78234

1) Expert system to assist in scheduling nurses for ward shifts.

2) Expert system to assist in Emergency Department disposition of chest pain patients.
3) Expert system to assist in Emergency Department triage of patients.

Expert System Name: FEVES

Purpose: Validate aircraft finite element models
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Nasir Munir,
Northrop Aircraft Division
1 Northrop Avenue
Hawthorne, CA 90250

Finite Element Validation Expert System aids in the automation of the analysis/design tasks in
the NASTRAN modeling of aircraft structures by addressing the problems of validation and
classification.

FEVES represents knowledge using frames and production rules. A frame based knowledge
representation is used for the classification portion and production rules are used to validate the
model. FEVES communicates with the user through a graphical interface. A forward chaining
mechanism is used to make expert decisions. The separate components of FEVES communicate
through a common database.

The chief benefits of using FEVES are: (1) automation of the errorprone task of classifying sub-
components in a large NASTRAN model. For large models the time saved is between 40 and 60
percent; (2) automation of the error checking process which leads to consistent modeling
practices. As a training vehicle, it aids inexperienced users to learn good modeling practices.

29
Expert System Name: None
Purpose: Resource allocation planning
Development Stage: Conceptual
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Michael McKenney
U.S. Army
75 SW Cutoff
Northboro, MA 01532

Given a current plan and a current state of resource allocation:

• Evaluate current state of plan against administrative rules.

• Evaluate current state against policy rules.
• Redistribute
• Evaluate new distribution against various other senarios.

Expert System Name: B2C2

Purpose: Distributed and stand-alone terminal application in support
of command and control (C2); decision support
Development Stage: Developing
Other Languages/Shells Used: KBEST/RBEST(TITAN) and M-1
Last Update: May 29, 1992
Contact: George E. Sherman
U.S. Army
ATTN: AMSEL-RD-C3-CC
Ft. Monmouth, NJ 07703

Utilize TAE+ for X Window user interface and C code generation, linked to TACTICIAN and
mapped data, distributed data access and position-location capability, supported by VHF and area
communications architectures. Map-based decision aid with task matrix format.

30
Expert System Name: GESLAN
Purpose: Allow choice of suitable conditions for laminar analysis
programs.
Development Stage: Developing
Other Languages/Shells Used: FORTRAN
Last Update: May 29, 1992
Contact: Michael K. Neylon
NASA Lewis Research Center
2100 Brookpark Rd.
MS - 49-8
Cleveland, OH 44135

GESLAN provides both an expert system and a direct input that creates the necessary files to run
the laminar programs.

Expert System Name: Intelligent tutoring system shell

Purpose: Provide a generic ITS Shell for Army personnel to use to
create ITS's for various areas.
Development Stage: Developing
Other Languages/Shells Used: C++
Last Update: May 29, 1992
Contact: Robert E. Scurlock
Naval Postgraduate School
108 Brownnell Circle
Monterey, CA 93940

Working on establishing generic modules of ITS design that can be put together with a
knowledge base to use as an ITS. Examining using C++ object interfaced with CLIPS 5.0.

Expert System Name: Intelligent Forecasting System

Purpose: PC-based forecasting workstation
Development Stage: Beta testing
Other Languages/Shells Used: Autobox 3.0, FORTRAN Forecasting Engine
Last Update: May 29, 1992
Contact: Nelson Marquina
KETRON
350 Technology Drive
Malvern, PA 19355-1315

Initial application provides GUI front end to a Box-Jenkins forecasting algorithm/engine. The
expert system will automatically interact with the forecasting engine to find better models for the
given time series. Application system can be used with time-domain or frequency-domain data.

31
Expert System Name: EVAL
Purpose: Incorporate Program Impact Advisor System into a
component module of Program Manager's Support System.
Development Stage: Conceptual
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Tanya Tran
EG&G Wasc Inc.
8809 Sudley Road
Manassas, VA 22110

The PIA currently exists as a fully functional prototype on a Xerox 1186 Workstation. The code
consists of roughly 4000 lines of INTERLISP-D. We are in the process of translating LISP into
CLIPS. Modification of CLIPS will be supplemented to interface with RDBMS
(Ingress/ORACLE/INFORMIX). The user interface will be needed later in our development.

Expert System Name: Value Engineering Change Proposal

Purpose: Calculating engineering change proposals
Development Stage: Finished
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Becky Williams,
U.S. Army
HQ AMCCOM, Attn: AMSMC-SAO
Rock Island, IL 61299-6000

When a contractor finds a less expensive way to provide a product the government shares the
savings with the contractor. The VECP program is an automated system which asks required
questions and then makes the required calculations.

Expert System Name: DMINS

Purpose: Diagnose faults in inertial guidance unit
Development Stage: Finished
Other Languages/Shells Used: KMS
Last Update: May 29, 1992
Contact: Eric Hanson
U.S. Air Force
WL/AAA-1
WPAFB, OH 45433

32
The user interface was done using KMS, a hypermedia system. The system runs on Sun 3 and
Sun 4 workstations.

Expert System Name: Post-Market Care

Purpose: Evaluate protocols for medical device studies.
Development Stage: Conceptual
Other Languages/Shells Used: SQL and C
Last Update: May 29, 1992
Contact: Robert Fike,
USFDA - Office of Science & Technology
MS - HFZ-142
12720 Twinbrook Parkway
Rockville, MD 20857

The Safe Medical Device Act of 1990 requires that certain types of medical devices be studied
after their introduction into the market. We hope to automate as much as possible the scientific
and administrative evaluations of the protocols for these post-market studies. This could include
accessing pre-market study information about a particular device from a relational database.
Also, the possibility of multiple linked expert systems exists.

Expert System Name: None

Purpose: militaryBattle manager for SDI simulation
Development Stage: Developing
Other Languages/Shells Used: FORTRAN and C
Last Update: May 29, 1992
Contact: Charles Bosch
General Electric
P. O. Box 1000
Blue Bell, PA 19422

Core simulation written in FORTRAN interfacing to CLIPS via C. C will call CLIPS with
various positional, situation, sensor data and will get “shot” strategy from a knowledge base.

33
Expert System Name: PNS
Purpose: Diagnosis of fluid dynamics analysis codes
Development Stage: Finished
Other Languages/Shells Used: ART, ART-IM, KEE, OPS5, Smalltalk, and Objective C
Last Update: May 29, 1992
Contact: James A. Simak
General Dynamics
3849 Misty Meadow Drive
Fort Worth, TX 76133

The Parabolized Navier-Stokes (PNS) Advisor is an embedded expert system application that
diagnoses and corrects problems encountered while running the AFWAL PNS Code. Originally
developed and running on a Cray XMP, the Advisor provided pre-execution analysis of a user's
problem specification and post-analysis of the solution and any error conditions that had
occurred. The Advisor currently runs on a Convex C240 and supports several CFD Analysis
including the AFWAL PNS Code.

An interactive knowledge acquisition tool was developed on an IRIS 4D/25 in 1990 to enable the
“export” developers to directly manage and modify the knowledge bases. A case library has been
added to provide extended diagnostics support for problems not explicitly identified in the
knowledge base. A case-base reasoning rulebase and automated case acquisition and indexing
methodology is under development.

Expert System Name: OPPLAN

Purpose: Operations Planning
Development Stage: Developing
Other Languages/Shells Used: LISP and MetaData
Last Update: May 29, 1992
Contact: Bruce H. Cottman,
Symbiotics, Inc.
60 Canterbury Road
Waltham, MA 02139

Integrate CLIPS with a logistics model based on an object-oriented DBMS.

34
Expert System Name: Payload Avionics Integration Tool
Purpose: Analysis of Shuttle payload electrical and avionic
requirements
Development Stage: Developing
Other Languages/Shells Used: ART, KEE, and Ada
Last Update: May 29, 1992
Contact: Robert Duane
Lockheed
P. O. Box 58561
Houston, TX 77258-8561

The tool is a relatively simple forward-chainer with a lot of comparison of input values to known
limits. A fairly sophisticated user interface has been written with Hypercard. The only interface
between the Hypercard application and the CLIPS application is through external text files.

Expert System Name: None

Purpose: Control and diagnosis of nuclear reactors
Development Stage: Developing
Other Languages/Shells Used: C, Objective C
Last Update: May 29, 1992
Contact: Luis Rovere
Oak Ridge National Laboratory
P. O. Box 2008
Oak Ridge, TN 37831-6010

Implementation of diagnostic rules to identify plant operational status and generate appropriate
control actions. Fuzzy control was included as a separate set of objects in Objective C. Now they
will be moved to COOL. These control systems are running on a Sun network using RPCS.

Expert System Name: TARGET v3.0

Purpose: Simulation and analysis of warfare engagements and systems
performance
Development Stage: Developing
Other Languages/Shells Used: C++
Last Update: May 29, 1992
Contact: Laura G. Hinton
Trident Systems Inc.
10201 Lee Highway, Suite 300
Fairfax, VA 22030

This project involves the complex development of a Tactical Resource Generation and
Evaluation Tool. The TARGET model is an object-oriented, Monti-Carlo simulation of multi-

35
warfare engagements that includes platforms (air, surface, and subsurface), sensors (platform-
based, and independent, such as buoys), weapons, C3I functions and tactics. Additional features
of the model include interaction with tactical procedures and operation through an embedded
expert system, and a sophisticated graphical user interface.

Expert System Name: Resource Management

Purpose: Develop scenarios for resource evaluation
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Dave DeMascio
General Electric
P. O. Box 8048
Philadelphia, PA 19101

Build knowledge bases to reuse previously developed applications knowledge; update the
knowledge; then use the newly modified knowledge to evaluate resources.

Expert System Name: FMES

Purpose: Assist field investigator at flight mishap site.
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Paul R. Cyrus,
U.S. Air Force
2049 C-CSG/SCDC
McClellan AFB, CA 95652

Flight Mishap Expert System allows the field investigator to collect data on a mishap
investigation. CLIPS is used to guide the investigator through the investigation. It utilitzes a
pen-based computer (GRIDPAD) to perform this function.

Expert System Name: None

Purpose: Meteorologic/Oceanographic forecasting
Development Stage: Conceptual
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Paul Ehrler,
Lockheed
6800 Burleson Road
O/T2-30 B/310
Austin, TX 78744

36
Initial plans are to develop a low-cost oceanographical work station using multi-source data an
visualization techniques. Plans are to upgrade to 386 or 486-based develpment with Turbo C++.
Decision to use CLIPS based on cost, growing user community, efficiency, suitability to
low-cost platforms.

Expert System Name: None

Purpose: aerospaceSupport human & habitation team for Lunar/Mars
program
Development Stage: Conceptual
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Martha Evert
Lockheed
2499 Nasa Road One
MC - C44
Houston, TX 77058

Currently reviewing CLIPS as a resource to be used in modeling a habitat on the Moon and
Martian surface. CLIPS would be used to make decisions on combining various subsystems
during an outpost operation. The decision-making would interface with Excel databases to
calculate subsystems’ mass, volume and power figures. Interested in obtaining HyperCLIPS as
another possibility. No formal presentation has been made to NASA. The team is reviewing
possibilities in parametric analysis.

Expert System Name: None

Purpose: Project management
Development Stage: Conceptual
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Gerald M. Powell
U.S. Army
HQ, U. S. Army, CECOM
Ft. Monmouth, NJ 07703

Specific applications would come from the Army's five battlefield functional areas, for example,
countability planning is a subfunction of the Maneuver control functional area.

37
Expert System Name: PreAmp
Purpose: PWA Producibility Advisor
Development Stage: Conceptual
Other Languages/Shells Used: C, LISP
Last Update: May 29, 1992
Contact: James C. Muller
Martin Marietta Electronics
P.O. Box 628007
12506 Lake Underhill, Dr.
Orlando, FL 32825

The PreAmp Program will significantly advance U.S. electronic product development. It will
demonstrate integration of engineering and manufacturing enabled by a modular,
standards-based data-sharing automation framework. The PreAmp program addresses the
technology, automation software, and information-sharing requirements that will enable the U.S.
electronics industry to obtain the huge competitive benefits integrated product development
studies have identified in limited, proprietary settings. Those benefits include a substantial
increase in product flexibility together with dramatic reductions in time-to-market, product
development cost and cost of quality. The data sharing and concurrent engineering automation
technology concepts developed in PreAmp will focus on the electrical/electronic domain. These
concepts will be demonstrated and technology barriers to wide-spread use will be identified.
Ultimately, this approach can be extended to other industries to obtain similar global competitive
advantages.

Expert System Name: None

Purpose: Simulation of distributed autonomous network management
algorithms & protocols
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Scott Wills,
Harris GASD
MS - 102-4844
P. O. Box 94000
Melbourne, FL 32902

Discrete event simulations of various network designs. Uses COOL extensively to describe
COMM components, uses rules to describe protocols & algorithms.

38
Expert System Name: HLPR
Purpose: Improve process by which optimal production methods and
procedures are forecast and determined
Development Stage: Conceptual
Other Languages/Shells Used: C, Vermont Views
Last Update: May 29, 1992
Contact: J. V. Dart,
LTV Aerospace & Defense
9314 W. Jefferson Blvd.
Dallas, TX

The system will automate tasks of pre-planning, bid-planning, tool ordering, electronic buyoff
and generation of various reports. The system is interactive and VAX-based using C and SQL.

Expert System Name: Eagle Analyst Workstation

Purpose: Detection of anomalies
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Dennis Powell
Los Alamos National Laboratory
MS - F602
Los Alamos, NM 87545

Apply expert system technology to reviewing the outputs of an Army combat model (Eagle) in
order to detect anomalies and subsequently point to potential errors in model input or algorithms.

Expert System Name: None

Purpose: Resource Planning
Development Stage: Finished
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Vincent Papandrea
General Dynamics
75 Eastern Point Rd.
D/449 C62
Groton, CT 06340-4989

Production flow through major facilities is modeled using input from individual currently
performing planning task (ie: the expert). System assigns jobs to appropriate work area in
accordance with end use date obtained from the company's planning databases
(MRP/ARTEMUS). Outputs include drawings of the facility showing location of units or ships at

39
any point in the schedule usage charts, and work center loading displays. The system has been
used extensively for analyzing facility requirements for a new product line.

Expert System Name: None

Purpose: JIT System for delivery of spare parts
Development Stage: Alpha Testing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Bill Spernow
Entelechy Research, Inc.
9448 Miranoy Drive
Sacromento, CA 95826-5227

Attempting to develop a CLIPS program that can be used as a structure, around which will be
built a JIT system for delivery of spare parts (from a central warehouse) used in maintenance and
repair of aircraft at an Air Force repair depot.

Expert System Name: Situation Assessment Subsystem

Purpose: Situation Assessment Subsystem for rotorcraft pilots
associate
Development Stage: Developing
Other Languages/Shells Used: C and C++
Last Update: May 29, 1992
Contact: Dan Ballaro
Reticular Systems, Inc.
4095 Calgary Ave
San Diego, CA 92122

The Situation Assessment Subsystem is being developed as a cognitive decision aid to assist the
pilot in high threat environments. The system uses multiple blackboards to recognize, analyze
and predict threat behavior and infer high level attributes about detected objects of interest.

Expert System Name: None

Purpose: Computer language conversion
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Robert L. Klungle,
Hughes Aircraft
506 N. Irena #B
Redondo Beach, CA 90277

40
• System for performing language conversion (FORTRAN, C, Pascal) to ADA. Also
improves ADA code currently running.
• Automatic generaion of SIMSCRIPT II.5 simulation code from graphic pictures.

Expert System Name: SIPS QA/DA

Purpose: aerospaceResolve errors during dowlink transmission from
Spacelab.
Development Stage: Finished
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: W. David Ripley III
Computer Sciences Corporation
4600 Powder Mill Road
Beltsville MD 20705

The Spacelab Input System (SIPS) Quality Assurance and Data Accountability (QA/DA) expert
system was developed for NASA to aid Spacelab Data Processing Facility (SLDPF) personnel.
The system assists SLDPF personnel in detecting and resolving anomalies that occur during a
downlink transmission from Spacelab and those that reesult from a variety of data processing
system failures associated with the capture and storage of large amounts of data.

Expert System Name: AIDBI

Purpose: Assist radar systems intelligence analysts
Development Stage: Conceptual
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Russell Moody,
ORION International Technologies, Inc.
4027 Colonel Glenn Highway: Suite 411
Dayton, OH 45431

The FASTC (Foreign Aerospace Science & Technology Center) computer environment consists
of large on-line databases (structured/formatted and free text), models and simulations, and other
analytical and data processing software tools. FASTC is transitioning from a PC to a Sun
client/server environment. The AIDBI program is investigating opportunities to apply expert
systems to improve the analysis process. Prime candidates which have been identified are:

1) Text search, retrieval and processing

2) Intelligence data research and collection
3) On-line radar systems expert advisor
4) Radar systems analyst tutor

41
The project is currently in the operational requirements definition stage.

Expert System Name: Rapidism

Purpose: Improve Rapidism transportation model decisions.
Development Stage: Conceptual
Other Languages/Shells Used: Ada
Last Update: May 29, 1992
Contact: Donald E. Clemmer
Computer Sciences Corporation
MS 312
3160 Fairview Park Drive
Falls Church, VA 22042

Purpose of concepts being worked out is to show government personnel how expert systems and
CLIPS can enhance model processing.

Expert System Name: Finance Studies

Purpose: Alternatives for manpower reductions
Development Stage: Conceptual
Other Languages/Shells Used: C, C++, XLisp, Clipper 5.0
Last Update: May 29, 1992
Contact: David T. Roberts
U.S. Army
HQ 3rd Corps at Fort Hood
Fort Hood, TX 76544-5056

We anticipate a number of retirees who won't be replaced during manpower reductions and
hiring freezes. We seek to use expert system to:

1) Capture their knowledge

2) Permit new employees to handle the most common problems without a lengthy training
period.
3) Use expert system as a training vehicle.

Areas under exploration: travel pay; estimating, travel pay settlement, allowable expenses;
commercial accounts payable; civilian pay queries; military pay queries; contract management;
interfund transfers.

Knowledge bases would be updated as regulations change. When old rules (superseded) are
overlooked and thus not deleted, conflict resolution becomes paramount.

42
Expert System Name: None
Purpose: Assist in combat simulation
Development Stage: Conceptual
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: H. M. Ryan III
U.S. Army
8120 Woodmont Ave
Bethesda, MD 20814

We are beginning a project in which we contemplate the use of CLIPS. The system will be rule
based and will compare and check the output of a combat simulation model against a database of
“clean” model runs. System purpose is to flag output elements that might appear not to be
representative of expected combat.

Expert System Name: AI/Ada Processor

Purpose: Improvement of Ada AI systems
Development Stage: Developing
Other Languages/Shells Used: Ada
Last Update: May 29, 1992
Contact: Bill Wavering
Integrated Software, Inc.
1945 Palm Bay Rd NE
Palm Bay, FL 32905

Particular application will be an expert system (pilot’s aid) for an en route mission planning
system. We are currently using an expert system written in Ada at Florida Institute of
Technology. We will be exploring the possibility of using our hardware to improve Ada CLIPS
performance.

Expert System Name: TBHELP

Purpose: Air to air combat simulation
Development Stage: Developing
Other Languages/Shells Used: FORTRAN 77
Last Update: May 29, 1992
Contact: Barry N. Cox
Lockheed
501 Marin St., Suite 214
Thousand Oaks, CA 91360

TAC BRAWLER consists of over 2,600 subroutines and 110,000 lines of code. It is an
exceptionally complex model because of the huge number of parameters which may be varied. It

43
is hoped that CLIPS may be embedded within the TBHelp Tool (an enhanced user interface tool)
to assist with failure diagnostics and reporting.

Expert System Name: Automated Message Update Module

Purpose: Integration of NASA data sources
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Glenn Oliveira
Marin Marietta
700 W. Mineral Ave
MS - XL 8058
Littleton, CO 80120

This would provide a front-end for diverse NASA data set header formats. Its functions would
include:

1) Integrate interdisciplinary data set information into a common format; and

2) Provide inference metadata to these input data sets based on prior data set knowledge. This
includes establishing a “confidence factor” for existing header information, and augmenting
existing header information with relevant user, site, interest domain, and storage
information.

Expert System Name: Intelligent Computer Aided Design System

Purpose: Use of expert systems in CAD
Development Stage: Finished
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: Jesse Allen
Accugraph Corporation
5822 Cromo Drive
El Paso, TX 79912-5598

Designing a new home or office building is much more than just transferring a mental image of
the structure onto a computer screen via your favorite CAD program. It involves the interaction
of an experienced architect and many specialists in areas as diverse as interior lighting and
structural engineering. The process of design, consultation with specialists, and revision can be a
long, arduous and costly affair.

The ICADS system consists of a number of expert systems developed with CLIPS and is
designed to run under Accugraph’s MountainTop CAD and Graphic Information Management

44
Software. Some of the experts that can be consulted include a structural analyst, lighting expert,
cost advisor, and an acoustical consultant.

Another expert system that Accugraph is using CLIPS and ICADS for is their “NetWork
Analyst” which also runs under MountainTop. It could be used to determine the connectivity of
objects in a network for fault analysis or rerouting.

Expert System Name: Financial Management System

Purpose: Financial data base system management
Development Stage: Conceptual
Other Languages/Shells Used: GURU and First Class
Last Update: May 29, 1992
Contact: Kerry Culligan
U.S. Navy
NCTS Code N813
Bldg 600
Pensacola, FL 32508

We have just become involved with the maintenance and enhancement of the Financial
Management System and the Inventory Management System. If applicable, we plan to combine
the two databases and use an expert system. One of these applications is written in GURU,
however, it appears to be mostly procedural code for the database.

Expert System Name: Risk Assessment Tool

Purpose: Provide an on-line risk assessment tool
Development Stage: Finished
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: David B. Ramsey,
PRC
600 Maryland Ave., S.W., Suite 400
Washington, DC 20024

This application provides the user with a sensitivity/criticality assessment.

45
Expert System Name: None
Purpose: Power supply contingency evaluation
Development Stage: Developing
Other Languages/Shells Used: None
Last Update: May 29, 1992
Contact: John Bremser,
Los Alamos National Laboratory
MS F661
Los Alamos, NM 87545

Given a particular power utility company and a description of their operating procedures, this
expert system can make recommendations as to actions to be taken in the event of a contingency
or contingencies.

46
 Index

Accugraph Corporation 44 Clipper 42

Ada 35, 42, 43 CommonKADS 6
aerospace 17, 22, 29, 35, 36, 37, 40, 41, 43 Compton, Michael 3
agriculture 23, 27 Computer Aided Aircraft design Package 4
AI/Ada Processor 43 Computer language conversion 40
AIAI 6, 7 Computer Sciences Corporation 41, 42
AIDBI 41 Control 35
Allen, Jesse 44 COOL 38
Ames Research Center 3 Cord, Thomas 5
Analysis 12, 35 Cottman, Bruce 34
anomaly detection 39 Coulson, Robert 23, 24, 25
AppleScript 3 COURSE SELECTOR 7
ART 34, 35 Cox, Barry 43
ART-IM 34 Cray Research 16
ASPEN 17 Culligan, Kerry 45
ASW Info Sys Dynamic Data Store 28 Cyrus, Paul 36
Autobox 31 Dart, J. V. 39
Automated Message Update Module 44 decision support 30
automation 5 DeMascio, Dave 36
AVL List GmbH 12 design 4, 17
B2C2 30 DI-3000 19
Baker, Nelson 13 diagnosis 11, 12, 32, 34, 35
Ballaro, Dan 40 Diagnostic & Maintenance 16
Bell Northern Research 15 Drive Analysis Expert System 11
blackboard 40 Duane, Robert 35
Blahowsky, Peter 12 Eagle Analyst Workstation 39
Bosch, Charles 33 EG&G Wasc Inc. 32
Bourne shell scripts 9, 10 Ehrler, Paul 36
Boyer, Douglas 27 EMI Detection Expert 23
BRAT 22 engineering 4, 12, 13, 19, 32
Bremser, John 46 Entelechy Research, Inc. 40
Brown, David 4 EPL 14
C 4, 5, 11, 12, 13, 33, 35, 38, 39, 40, 42 Esch, John 28
C++ 5, 31, 35, 40, 42 ET++ 5
C-shell 11 EVAL 32
CAAP 4 Event Pattern Language 14
CAD 44 Evert, Martha 37
CAD/CAM 18 Excel 37
cancer patients 28 FASTC 41
Clemmer, Donald 42 FEVES 29
Cleveland State University 15 Fike, Robert 33

47
file storage 28 Informed Manager 3
FileMaker Pro 3 INSEX 24
Finance Studies 42 Integrated Software, Inc. 43
Financial 45 Intelligent Computer Aided Design System 44
Financial Management System 45 Intelligent Forecasting System 31
FIRE-XPS 12 Intelligent Remote Automation Project 5
First Class 45 Intelligent Tutoring System 13
Flight Mishap Expert System 36 Intelligent tutoring system shell 31
Florida Institute of Technology 43 Iowa State University 16
FMES 36 ISPPEX 25
food processing 5 ITS-Engineering 13
forecasting 31 JIT System 40
forms 3 KBEST/RBEST 30
Forschungszentrum Informatik 5 KEE 34, 35
FORTRAN 12, 19, 31, 33, 43 KETRON 31
G2 5 Kingston, John 6, 7
General Dynamics 34, 39 Klungle, Robert 40
General Electric 16, 33, 36 KMS 32
GermAlert 10 knowledge acquisition 6
GermWatcher 9 Koval, Albert 16
GESLAN 31 Lamiell, James 29
GIS 27 Langley Research Center 19, 21
GIS, 24 Laser Docking Sensor Associate 17
Giuffrida, Giovanni 14 Leukemia Diagnostic Assistant 28
Given, Mott 26 Level5 5
Gold, Steven 16 Lewandowski, Henry 15
Gossage, Brett 22 Lewis Research Center 15, 31
GRASS 24 LISP 34, 38
Green, Peter 27 Lockheed 28, 35, 36, 37, 43
Griffith, Paul 17, 18 Los Alamos National Lab 17, 28
GURU 45 Los Alamos National Laboratory 39, 46
Hanson, Eric 32 LTV Aerospace & Defense 39
HARDY 6 M-1 30
Harris GASD 38 MacArthur, Robert 27
Helland, Barbara 16 Macintosh Toolbox 4
Hinton, Laura 35 Mack Technologies, Inc. 11
HLPR 39 manpower reductions 42
Hughes Aircraft 40 Marin Marietta 44
Hypercard 35 Marquina, Nelson 31
HyperCLIPS 37 Martin Marietta Electronics 38
Hypertext 6 McKenney, Michael 30
ICADS 44 medical 9, 10, 33
infection control 9, 10 Medical expert systems 29

48
MetaData 34 Powell, Gerald 37
Meteorologic/Oceanographic forecasting 36 PowerShare 3
Mettrey, William 15 PowerTalk 3
Microexpert Systems, Inc. 17, 18 PRC 45
Microsoft Windows 7 PreAmp 38
military 33, 35, 39, 43 Price quotation 15
model processing 42 Princeton University 4
monitoring 15 PRISM 15
Moody, Russell 41 Project management 37
Moore, Greg 11 PROLOG 19
Muller, James 38 PROMEAT 5
multimedia 13 propulsion systems 15
Munir, Nasir 29 Prototype Electronic Purchase Request 3
Murphey, Ann 22 Quality inspection 5
NASA 3, 15, 19, 21, 31, 37, 41, 44 radar systems 41
Naval Postgraduate School 31 Ramsey, David 45
network 4 Rapidism 42
NetWork Analyst 45 Resource Management 36
network management 38 Reticular Systems, Inc. 40
Neylon, Michael 31 RIM 19
Nichols Research Corporation 22 Ripley, David 41
Northrop Aircraft Division 29 risk assessment 45
Numerical Propulsion System Simulation Risk Assessment Tool 45
(NPSS) Monitor 15 Roberts, David 42
Oak Ridge National Laboratory 35 Rockwell Space Operations Co. 22
Objective C 34, 35 Rogers, James 19, 21
Oliveira, Glenn 44 Rovere, Luis 35
OPPLAN 34 RPC 12
OPS5 34 Ryan, H. M. 43
optimization 39 Salzman, Gary 28
ORION International Technologies, Inc. 41 Schaller, Rob 28
Papandrea, Vincent 39 Scurlock, Robert 31
Paramax Systems Inc. 28 SEI 23
Parkinson.i., W.J. 17 SelectPC 28
Payload Avionics Integration Tool 35 Sherman, George 30
PEPR 3 Simak, James 34
Planners Workbench 18 simulation 15, 22, 33, 35, 38, 43
planning 18, 30, 34, 39, 43, 46 SIPS QA/DA 41
Planning and scheduling 16 Situation Assessment Subsystem 40
PNS 34 SLATEC 16
Polycenter NetView 4 Smalltalk 34
Post-Market Care 33 sockets 12
Powell, Dennis 39 Solaris 12

49
Southern Pine Beetle 24 Wills, Scott 38
Spacelab 41 WindExS 5
Spernow, Bill 40 Woodpecker 23
SQL 33, 39 Worcester Polytechnic Institute 4
Steib, Sherry 9, 10 Worchester Polytechnic Institute 27
STRUTEX 19 X Window 30
SunOS 12 X Windows 7
Swanson, Dave 23 XLisp 42
Sybase 9, 10, 14 XVT 4
Symbiotics, Inc. 34 Yalif, Guy 4
TAC BRAWLER 43
TAE+ 13, 30
TARGET 35
TBHELP 43
temporal reasoning 14
TENNIS 4
Texaco Trading & Transportation, Inc. 5
Texas A&M University 23, 24, 25
The Open Practical Knowledge Acquisition
Toolkit 6
TOPKAT 6
Toupin, Edward 5
Tran, Tanya 32
Trident Systems Inc. 35
TSR Graphics 17
U.S. Air Force 32, 36, 40
U.S. Army 22, 26, 30, 32, 37, 39, 42, 43
U.S. Navy 23, 45
UCLA 14
University of Arizona 27
Unix 12, 26
UNIX Advisor 26
USDA 27
USFDA 33
Value Engineering Change Proposal 32
VAX 39
Verification and validation 27
Vermont Views 39
Washington University School of Medicine 9,
10
Wavering, Bill 43
Weed Control Advisor 23
Williams, Becky 32

50