CPE CSC 481 Knowledge-Based Systems Dr Franz J Kurfess

Usage of the Slides u these slides are intended for the students of my CPE/CSC 481 “Knowledge-Based Systems” class at Cal Poly SLO u if you want to use them outside of my class, please let me know (fkurfess@calpoly. edu) u. I usually put together a subset for each quarter as a “Custom Show” u u To to view these, go to “Slide Show => Custom Shows”, select the respective quarter, and click on “Show” print them, I suggest to use the “Handout” option 4, 6, or 9 per page works fine u Black & White should be fine; there are few diagrams where color is important u © 2002 -9 Franz J. Kurfess Expert System Examples 2

Course Overview u Introduction u Knowledge u with Uncertainty Probability, Bayesian Decision Making u Expert u and Inference Predicate Logic, Inference Methods, Resolution u Reasoning u Representation Semantic Nets, Frames, Logic u Reasoning u u CLIPS System Design u Overview Concepts, Notation, Usage u Pattern u Matching Variables, Functions, Expressions, Constraints u Expert System Implementation u Salience, Rete Algorithm u Expert System Examples u Conclusions and Outlook ES Life Cycle © 2002 -9 Franz J. Kurfess Expert System Examples 3

Overview ES Examples u Motivation u Objectives u Chapter u u u Introduction Review of relevant concepts Overview new topics Terminology u R 1/XCON u u u System Configuration Knowledge Representation Reasoning u Human Resources ES u OSHA Hazard Awareness Advisor u Gensym G 2 Real-Time Expert System u Important Concepts and Terms u Chapter Summary u MYCIN © 2002 -9 Franz J. Kurfess Expert System Examples 4

Logistics u u Introductions Course Materials u u textbooks (see below) lecture notes u u u handouts Web page u u u Power. Point Slides will be available on my Web page http: //www. csc. calpoly. edu/~fkurfess Term Project Lab and Homework Assignments Exams Grading © 2002 -9 Franz J. Kurfess Expert System Examples 5

Bridge-In © 2002 -9 Franz J. Kurfess Expert System Examples 6

Pre-Test © 2002 -9 Franz J. Kurfess Expert System Examples 7

Motivation u reasons to study the concepts and methods in the chapter u main advantages u potential benefits u understanding of the concepts and methods u relationships to other topics in the same or related courses © 2002 -9 Franz J. Kurfess Expert System Examples 8

Objectives u regurgitate u basic facts and concepts u understand u u elementary methods more advanced methods scenarios and applications for those methods important characteristics v differences between methods, advantages, disadvantages, performance, typical scenarios u evaluate u application of methods to scenarios or tasks u apply u methods to simple problems © 2002 -9 Franz J. Kurfess Expert System Examples 9

R 1/XCON u one of the first commercially successful expert systems u application domain u u u configuration of minicomputer systems selection of components arrangement of components into modules and cases u approach u u data-driven, forward chaining consists of about 10, 000 rules written in OPS 5 u results u u u quality of solutions similar to or better than human experts roughly ten times faster (2 vs. 25 minutes) estimated savings $25 million/year © 2002 -9 Franz J. Kurfess Expert System Examples 11

System Configuration u complexity u u components u u u tens or hundreds of components that can be arranged in a multitude of ways in theory, an exponential problem in practice many solutions ``don't make sense'', but there is still a substantial number of possibilities important properties of individual components stored in a data base constraints u functional constraints derived from the functions a component performs v u e. g. CPU, memory, I/O controller, disks, tapes non-functional constraints v such as spatial arrangement, power consumption, © 2002 -9 Franz J. Kurfess Expert System Examples 12

Knowledge Representation u configuration space u constructed incrementally by adding more and more components v the correctness of a solution often can only be assessed after it is fully configured u subtasks v are identified make the overall configuration space more manageable u component u retrieved u control u rules from the external data base as needed knowledge that govern the sequence of subtasks u constraint u rules knowledge that describe properties of partial configurations © 2002 -9 Franz J. Kurfess Expert System Examples 13

Example Component u u partial description of RK 611* disk controller facts are retrieved from the data base and then stored in templates RK 611* Class: Type: Supported: Priority Level: Transfer Rate: . . . © 2002 -9 Franz J. Kurfess Uni. Bus module disk drive yes buffered NPR 212 Expert System Examples 14

Example Rule u rules incorporate expertise from configuration experts, assembly technicians, hardware designers, customer service, etc. Distribute-MB-Devices-3 If the most current active context is distributing Massbus devices & there is a single port disk drive that has not been assigned to a Massbus & there are no unassigned dual port disk drives & the number of devices that each Massbus should support is known & there is a Massbus that has been assigned at least one disk drive and that should support additional disk drives & the type of cable needed to connect the disk drive to the previous device is known Then assign the disk drive to the Massbus © 2002 -9 Franz J. Kurfess Expert System Examples 15

Configuration Task u u u u check order; identify and correct omissions, errors configure CPU; arrange components in the CPU cabinet configure Uni. Bus modules; put modules into boxes, and boxes into expansion cabinets configure panels; assign panels to cabinets and associate panels with modules generate floor plan; group components and devices determine cabling; select cable types and calculate distances between components this set of subtasks and its ordering is based on expert experience with manual configurations © 2002 -9 Franz J. Kurfess Expert System Examples 16

Reasoning u data-driven (forward chaining) u components are specified by the customer/sales person u identify a configuration that combines the selected components into a functioning system u pattern matching u activates u execution appropriate rules for particular situations control ua substantial portion of the rules are used to determine what to do next u groups of rules are arranged into subtasks © 2002 -9 Franz J. Kurfess Expert System Examples 17

Performance Evaluation u notoriously difficult for expert systems u evaluation criteria u usually very difficult to define u sometimes comparison with human experts is used u empirical evaluation u Does the system perform the task satisfactorily? u Are the users/customers reasonably happy with it? u benefits u faster, fewer errors, better availability, preservation of knowledge, distribution of knowledge, etc. u often based on estimates © 2002 -9 Franz J. Kurfess Expert System Examples 18

Development of R 1/XCON u R 1 prototype u the initial prototype was developed by Carnegie Mellon University for DEC u XCON commercial system u used for the configuration of various minicomputer system families u first VAX 11/780, then VAX 11/750, then other systems u reimplementation more systematic approach to the description of control knowledge v clean-up of the knowledge base v performance improvements v © 2002 -9 Franz J. Kurfess Expert System Examples 19

Extension of R 1/XCON u addition u u wider class of data additional computer system families u new u u u of new knowledge components refined subtasks more detailed descriptions of subtasks revised descriptions for performance or systematicity reasons u extended u configuration of ``clusters'' v tightly interconnected multiple CPUs u related u task definition system XSEL tool for sales support © 2002 -9 Franz J. Kurfess Expert System Examples 20

Summary R 1/XCON u commercial u u success after initial reservations within the company, the system was fully accepted and integrated into the company's operation widely cited as one of the first commercial expert systems u domain-specific u the availability of enough knowledge about what to do next was critical for the performance and eventual success of the system u suitability u u control knowledge of rule-based systems appropriate vehicle for the encoding of expert knowledge subject to a good selection of application domain and task © 2002 -9 Franz J. Kurfess Expert System Examples 21

MYCIN based on a presentation by Adam Gray, CSC 481 W 04 some modifications

Overview u History u DENDRAL u MYCIN u Background u Knowledge Representation u Knowledge Manipulation u Uncertainty u Performance Evaluation u Advantages and Problems u References © 2002 -9 Franz J. Kurfess A. Gray, 2004 Expert System Examples 23

DENDRAL u u Commonly considered the first expert system Developed at Stanford in the late 1960 s u u Ed Feigenbaum (a CSC Professor) Bruce Buchanan (a philosopher turned computer scientist) Joshua Lederberg (a Nobel Laureate Geneticist) Analyzed NMR mass spectrogram data to determine the geometric arrangement of atoms in a molecule © 2002 -9 Franz J. Kurfess A. Gray, 2004 Expert System Examples 24

MYCIN Background u Medical expert system u u Developed at Stanford in the 1970 s by Feigenbaum, Buchanan and Ted Shortliffe (a doctor) Recommended therapy for blood/meningitis infections v v u the diagnosis normally involved growing cultures of the infecting organism (48 hours) Doctors had to come up with quick guesses about likely problems v Prescribe drugs to deal with immediate problems Developed to explore how doctors make these rough, but important, guesses with partial information v Also important in practice as there are many junior doctors or nonspecialized doctors © 2002 -9 Franz J. Kurfess A. Gray, 2004 Expert System Examples 25

MYCIN Implementation u Goal-directed system that uses a basic backwardchaining technique u 450 Rules written in LISP u Performed as well as some experts and significantly better than junior doctors u Never actually used in practice u Not due to its performance u But rather ethical and legal issues © 2002 -9 Franz J. Kurfess A. Gray, 2004 Expert System Examples 26

Example Rule u If u The site of the culture is blood u The gram of the organism is neg u The morphology of the organism is rod u The burn of the patient is serious Then there is weakly suggestive evidence (0. 4) that the identity of the organism is pseudomonas © 2002 -9 Franz J. Kurfess A. Gray, 2004 Expert System Examples 27

Representation u Rules had no variables, contexts instead u u u MYCIN dealt with a number of implicit variables For example there could be a patient, a culture, a few infectious organisms. MYCIN’s knowledge structured into “object-parameter -value” triples u u u “culture” would be an object “site” would be a parameter of “culture” a possible value of this parameter would be “blood” © 2002 -9 Franz J. Kurfess A. Gray, 2004 Expert System Examples 28

Manipulation u MYCIN starts out with a rule that says u If there is an organism requiring therapy, then, compute the possible therapies and pick the best one u First u if tries to see if the disease is known it isn’t begins reasoning process u Basic routine in MYCIN u attempt to find the value of a parameter © 2002 -9 Franz J. Kurfess A. Gray, 2004 Expert System Examples 29

Finding Values u u Depending on the type of data may ask user if the value is known Tried to ask the most general question possible, so as not to become annoying or repetitive u u E. g. , if MYCIN wants to know if morphology of organism is rod, will ask “What is morphology of organism? ” rather than a specific question repeatedly Format of KR is supposed to make questions reasonable u If the value is not know, MYCIN does backward chaining u Stores a list of rules that might yield a value for each parameter © 2002 -9 Franz J. Kurfess A. Gray, 2004 Expert System Examples 30

Uncertainty u Medical field must reason in the presence of unknown, incomplete, vague or uncertain information u MYCIN used “certainty factors” initially hard to defend from a sound theoretical viewpoint v theoretical foundations were established later (Dempster-Shafer) v useful to see where knowledge about uncertainty exists, and the implications it has for the design of the system v © 2002 -9 Franz J. Kurfess A. Gray, 2004 Expert System Examples 31

Certainty Factors u u Range from – 1 (positive it is not the case) to +1 (positive it is the case). MYCIN maintains certainty for u possible values of parameters (ultimately, the certainty that you have a particular disease) v u u can maintain multiple possible values, each with its own certainty validity of a rule MYCIN has rules for combining the certainty factors © 2002 -9 Franz J. Kurfess A. Gray, 2004 Expert System Examples 32

Performance Evaluation u Shortliffe used u 10 sample problems u 8 otherapy recommenders v 5 faculty at Stanford Med. School, 1 senior resident, 1 senior postdoctoral researcher, 1 senior student u 8 impartial judges gave 1 point per problem u Max score was 80 u MYCIN: 65, Faculty: 40 -60, Fellow: 60, Resident: 45, Student: 30 © 2002 -9 Franz J. Kurfess A. Gray, 2004 Expert System Examples 33

Controls u Judges’ bias for/against computers u u Difficulty of problems u u Judges did not know who recommended each therapy Medical student did badly, so problems not easy Level of Interest u u Hypothesis in MYCIN that “knowledge is power” Have groups with different levels of knowledge © 2002 -9 Franz J. Kurfess A. Gray, 2004 Expert System Examples 34

Good Points u MYCIN was good in that u It could calculate dosages very precisely u Dealt well with interactions between drugs v An area in which humans have trouble u Possesses v nice explanation facilities Retrieves and displays relevant rules to offer explanation of its behavior © 2002 -9 Franz J. Kurfess A. Gray, 2004 Expert System Examples 35

Difficulties u Narrow u did in scope not scale up well to larger problems u Practical concerns u Doctors have reservations about advise from computers u Legal issues © 2002 -9 Franz J. Kurfess A. Gray, 2004 Expert System Examples 36

References l l l E. H. Shortliffe, F. S. Rhame, S. G. Axline, S. N. Cohen, B. G. Buchanan, R. Davis, A. C. Scott, R. Chavez-Pardo, & W. J. van Melle. “MYCIN: A computer program providing antimicrobial therapy recommendations”. Clinical Medicine, (Issue): 34, 1975. E. H. Shortliffe. “MYCIN: A rule-based computer program for advising physicians regarding antimicrobial therapy selection”. Proceedings of the ACM National Congress (SIGBIO Session), 739. 1974. Giarratano, J. and G. Riley, ``Expert Systems – Principles and Programming'' 3 rd Edition, PWS Publishing Company, 1998. “MYCIN: A Quick Case Study”. <http: //www. cee. hw. ac. uk/~alison/ai 3 notes/section 2_5_5. html>. Russel, Stuart J. and Peter Norvig. “Artificial Intelligence, A Modern Approach”. Prentice -Hall, Inc. , 2003. © 2002 -9 Franz J. Kurfess A. Gray, 2004 Expert System Examples 37

Human Resources Expert System u expert systems to determine conditions and entitlements for public employees in New South Wales, Australia u main user groups u u u employees managers HR staff u accessible u u via Internet http: //www. premiers. nsw. gov. au/Work. And. Business/Working. For. Govern ment/HRExpert. htm some functionality limited to authorized users u developed by Softlaw Corporation http: //www. softlaw. com. au © 2002 -9 Franz J. Kurfess Expert System Examples 38

Objectives u improve HR advice and information u quality, u enable u e. g. consistency, timeliness value-adding strategic functions work force planning u extend use of technology from transaction-based ES to advice and information systems © 2002 -9 Franz J. Kurfess Expert System Examples 39

HR Expert Principles u enhanced electronic decision tree u on-line inquiries from users determine branches u accessible via official HR web sites u integrated with source documents u legislation, © 2002 -9 Franz J. Kurfess personnel handbook, etc. Expert System Examples 40

HR Expert: Inquiries © 2002 -9 Franz J. Kurfess Expert System Examples 41

HR Expert: Service History © 2002 -9 Franz J. Kurfess Expert System Examples 42

Output u summary screens u reports u letters u applications u audit and forms

HR Expert: Summary Report © 2002 -9 Franz J. Kurfess Expert System Examples 44

HR Expert: Full Report © 2002 -9 Franz J. Kurfess Expert System Examples 45

Project u phase u u 3 agencies, 1, 250 staff, conducted in 2002 -3 demonstrated potential savings, user satisfaction, qualitative benefits u phase u u u 1 - pilot project 2 extension to all relevant conditions and entitlements to be operational by May 2004 cited in the report of the Australian Government - Information Management office as an example u technology u legislative rulebase technology, STATUTE Expert, by Softlaw Corp. , Canberra, Australia, http: //www. softlaw. com. au © 2002 -9 Franz J. Kurfess Expert System Examples 46

Benefits u employees u u u immediate and up to date information about conditions and entitlements easy access for inquiries improved data for decisions increased equity on-demand generation of reports standardized outputs and audit reports u Human u u v u u © 2002 -9 Franz J. Kurfess direct access to information about entitlements less tedious work v u Resources e. g. looking up information when employees need it reduced need for repetitive work more consistent decisions on-demand generation of reports standardized reports Expert System Examples 47

Issues u some u not u only of the input provided by the users always accurate, up to date generic conditions and entitlements u special u limited u not cases not included coverage all laws and regulations included u requires computer and Web access u commitment and buy-in from staff and employees © 2002 -9 Franz J. Kurfess Expert System Examples 48

Status u operational u update to include recent changes in laws and regulations under way u current u u u and in use modules Maternity Leave Study Time Extended Leave Recognition of Previous Service Leaving the Service Voluntary Redundancy Travel Compensation, including Meal and Private Motor Vehicle Allowances Higher Duties A llowance Salary Packaging Agency List Inquiry © 2002 -9 Franz J. Kurfess Expert System Examples 49

Rule. Burst Demo u a Flash demo of the Rule. Burst enviroment is available at http: //www. ruleburst. com/uploads/files/Rule. Burst. html u a predecessor of Rule. Burst was used to develop the HR Expert application © 2002 -9 Franz J. Kurfess Expert System Examples 50

References u u u HR Expert Case Study at http: //www. agimo. gov. au/resources/ppt/2003/030926 sb NSW governmental web site at http: //www. premiers. nsw. gov. au/Work. And. Business/Working. For. Governme nt/HRExpert. htm Australian Government Information Management Office Report at http: //www. agimo. gov. au/publications/2004/05/egovt_challenges/accounta bility/determinations/conclusion Softlaw Corporation Web site http: //www. softlaw. com. au Softlaw HR Expert Announcement http: //www. softlaw. com. au/content. cfm? categoryid=12&topicid=49&infopa geid=152 Rule. Burst KB Development Environment http: //www. ruleburst. com/ sites visited 03 -02 -05, 02 -28 -06 © 2002 -9 Franz J. Kurfess Expert System Examples 51

OSHA Hazard Awareness Advisor u asks questions about workplace activities, equipment, materials u analyzes the user’s answers u generates a report with common occupational hazards, applicable OSHA standards, and contacts u developed by the U. S. Department of Labor, Occupational Safety & Health Administration (OSHA) u version 1. 0 released in September 1999 u http: //www. osha. gov/dts/osta/oshasoft/hazexp. html © 2002 -9 Franz J. Kurfess Expert System Examples 52

Objectives u to help identify and understand common safety and health hazards in the work place u especially u designed u may aimed at small businesses for beginners be useful for experts as well u widely available through online and downloadable versions u downloadable © 2002 -9 Franz J. Kurfess version only for MS Windows Expert System Examples 53

Limitations u may not identify all hazards u will not determine compliance with OSHA standards u not a substitute for safety and health professionals © 2002 -9 Franz J. Kurfess Expert System Examples 54

OSHA Principles u expert system technology u accessible via official OSHA web sites u integrated with source documents u standards, © 2002 -9 Franz J. Kurfess legislation, etc. Expert System Examples 55

Hazard Awareness Advisor Inquiries © 2002 -9 Franz J. Kurfess Expert System Examples 56

Hazard Awareness Advisor Report u highlights u details © 2002 -9 Franz J. Kurfess

Report Highlights + Evaluate the exposure to chemicals in your workplace. + Your site needs a hazard communication program. + Inspect ladders and ensure that workers know how to use them safely. + Ladders should be at least 3 feet higher than the level they are going to reach. + Check that fire exits are unlocked and numerous enough for quick escape of all workers. + Keep passageways clear of obstructions. + Evaluate personal protective equipment that your workers purchase for themselves. + Protective eye wear may be needed to protect against splashes and sprays. + Use laser pointers carefully. They can cause eye damage. + Please investigate the need for head protection. + Please investigate the need for hard toed shoes. + Protective gloves may be needed because of injuries by knives or other hand tools. + Portable fire extinguishers must have maintenance service at least once a year and a written record must be kept to show the maintenance or recharge date. + Mark fuse boxes or breaker boxes to identify the circuits or equipment they control. + Extension cords should not be used as a substitute for permanent wiring. + Take care when using cleaning solvents and liquids when cleaning inside a series of deep cabinets or similar spaces. + If you have both ammonia and bleach cleaners, take care in their storage and use. The mixing of ammonia and bleach can produce dangerous chlorine gas. © 2002 -9 Franz J. Kurfess Expert System Examples 58

OSHA Report Detail: Portable Ladders Your answers indicate that your workers use portable ladders. The use of any ladder is hazardous. Workers may fall from them, fall with them, be struck by falling ladders or struck by objects dropped from work being performed on the ladder. Injuries also result from poor ladder placement: unstable footing, work angle too steep or too shallow, or placement in front of doors or passageways. Many serious falls from ladders are the result of workers standing above the designed working height of the ladder. The hazards of ladder use can be reduced by careful selection of ladders of appropriate height and strength, by routine inspection and maintenance, and by training of workers in safe ladder use. In order to safely gain access to an upper level such as a roof or platform, the portable or extension ladder must extend at least 3 feet above the point of contact. Any portable ladders should be tied off or held in position during use. © 2002 -9 Franz J. Kurfess Expert System Examples 59

Project u preceded u u Asbestos in '95 Confined Spaces in '96 Fire Safety in '97 Lead in Construction in '98 u input u u u u by other SHA advisors from National Federation of Independent Business, National Apartment Association, Synthetic and Organic Chemical Manufacturers Association, United Brotherhood of Carpenters and Joiners, Laborers Safety and Health Fund, International Brotherhood of Teamsters, other industry and labor organizations © 2002 -9 Franz J. Kurfess Expert System Examples 60

Benefits u owners/managers u u u easy access for inquiries about potential hazards quick analysis of the workplace generation of a report with highlights, details, and pointers for further information © 2002 -9 Franz J. Kurfess u employees u u u identification of potential hazards improved working conditions possible compliance issues v standards and regulations Expert System Examples 61

Issues u all of the input provided by the users u not always accurate,

Status u version u 1. 0 operational and in use since September 1999 update

Other OSHA Expert Advisors u see u u u u u http: //www. osha. gov/dts/osta/oshasoft/index. html Asbestos Confined Space Electronic Permit Required Confined Spaces (e-PRCS) Electronic Health and Safety Plan (e-HASP) Fire Safety Hazard Awareness Lead in Construction Lead in General Industry Lockout/Tagout v u u LOTO Plus Safe. Care $afety Pays © 2002 -9 Franz J. Kurfess Expert System Examples 64

References u u u OSHA Hazard Awareness Advisor, Version 1. 0 September 1999, http: //www. osha. gov/dts/osta/oshasoft/hazexp. html Stern, Ed (1998): “OSHA Unveils Online Hazard Awareness Advisor”, Access America Government Services, http: //govinfo. library. unt. edu/accessamerica/docs/expertadvisor. html Stern, Ed (1999): “The OSHA Hazard Awareness Advisor”, PC AI Magazine, vol 13, no 2, March/April 1999 Shirley, Robin E. (200): “New OSHA Interactive Software Designed to Help Small Business Owners”, On Target - News for the Small Business Owner, http: //www. reswritingservices. com/osha. html Virginia Workers’ Compensation Program (2005): “Hazard Assessments”, http: //www. covwc. com/lcarticles/archives/000084. php OSHA e. Tools and Electronic Products for Compliance Assistance http: //www. osha. gov/dts/osta/oshasoft/index. html sites visited 07 -14 -05, 02 -28 -06 © 2002 -9 Franz J. Kurfess Expert System Examples 65

Gensym G 2 u real-time expert system u developed by Gensym Corporation http: //www. gensym. com u application areas u chemical, oil & gas, process manufacturing, discrete manufacturing, power utilities, water utilities, telecommunications, government, transportation, aerospace u augmented © 2002 -9 Franz J. Kurfess by additional modules Expert System Examples 66

Gensym G 2 Platform © 2002 -9 Franz J. Kurfess Expert System Examples 67

Gensym G 2 Use http: //www. gensym. com/images/pages/g 2 platform. jpg © 2002 -9

Gensym Development Cycle http: //www. gensym. com/images/pages/xtreme-programming-lifecycle. jpg © 2002 -9 Franz J. Kurfess

Ericsson Network Management System u use of G 2 for wireless network management u challenges in wireless networks u additional v functionality instant messaging, chat, Web access, photos, videos, … u increased size and complexity of the network u very rapid growth and change rate u network management can be very stressful u constant v stream of alarms not all are important u extreme pressure to identify and fix problems © 2002 -9 Franz J. Kurfess Expert System Examples 70

Ericsson FMX u largely u u concentration on the fault management process reacts to all identified events very quickly v v v u u u automated system for wireless network management much faster than humans more reliable but less flexible filters out unimportant messages allows network operators to concentrate on critical events consolidates information for critical events u manages u u over 500, 000 events per day 500 systems in 100 countries 50 different equipment types u benefits u u increased quality of service reduced operating expenses © 2002 -9 Franz J. Kurfess Expert System Examples 71

FMX Screenshot Input Condition Decision © 2002 -9 Franz J. Kurfess http: //www. gensym.

Dow Chemicals Closed Loop Optimizer u energy management in a large petrochemical plant in Seadrift, TX u highly interdependent systems u real-time control u safety-critical u very high energy costs u utilization of waste heat from gas turbines u internal energy usage u excess energy sold to the grid u manual control of power generation was problematic u trade-off considerations must be made very fast © 2002 -9 Franz J. Kurfess Expert System Examples 73

Energy Management with G 2 u modeling of the energy system u sensors provide input u important components are modeled u output controls actuators, informs operators u previous models of individual systems were not successful for the overall energy management u optimization u determines the best operational plan for the current conditions in real time © 2002 -9 Franz J. Kurfess Expert System Examples 74

Seadrift Functional Diagram © 2002 -9 Franz J. Kurfess Expert System Examples 75

Seadrift Screenshot © 2002 -9 Franz J. Kurfess Expert System Examples 76

Seadrift Outcome u plant ran in closed loop mode 98 percent of the time u saved Dow $1. 25 million dollars in energy costs over one year u even larger potential for savings extension to other components and systems in the plant v more sophisticated modeling v usage for other plans v u user satisfaction u operators were skeptical initially, but accepted and used the system very quickly u better view of overall plant operations © 2002 -9 Franz J. Kurfess Expert System Examples 77

References u Gensym Corporation http: //www. gensym. com u Gensym “Success Stories”: Dow Chemicals http: //www. gensym. com/? p=success_stories&id=8 u Dow Chemicals Seadrift Plant http: //www. dow. com/ucc/locations/seadrift/about/inde x. htm © 2002 -9 Franz J. Kurfess Expert System Examples 78

Questions © 2002 -9 Franz J. Kurfess Expert System Examples 79

Figure Example © 2002 -9 Franz J. Kurfess Expert System Examples 80

Post-Test © 2002 -9 Franz J. Kurfess Expert System Examples 81

Important Concepts and Terms u u u u agenda backward chaining common-sense knowledge conflict resolution expert system (ES) expert system shell explanation forward chaining inference mechanism If-Then rules knowledge acquisition © 2002 -9 Franz J. Kurfess u u u knowledge base knowledge-based system knowledge representation Markov algorithm matching Post production system problem domain production rules reasoning RETE algorithm rule working memory Expert System Examples 83