This entry is missingOverview, add relevant content to make the entry more complete, and it can also be upgraded quickly. Hurry upeditCome on!
Expert system is an intelligent computerprogramSystem, which contains a large number ofDomain expertsLevel of knowledge and experience, which can apply artificial intelligence technology andcomputer technologyAccording to the knowledge and experience in the system, reasoning and judgment are carried out to simulate thedecision-making processIn order to solve the problems that need human experts to deal withComplex problemsIn short, expert system is a computer program system that simulates human experts to solve domain problems.
Expert system is one of the most important and active application fields in AI. It has achieved a major breakthrough in AI from theoretical research to practical application, from general reasoning strategy discussion to the application of specialized knowledge.Expert system is an important branch of early artificial intelligence, which can be regarded as a kind of computer with special knowledge and experienceSmart ProgramsThe system generally adopts theknowledge representation andKnowledge reasoningTechnology is usually simulated byDomain expertsCan be solvedComplex problems。
Origin and development
Announce
edit
expert system
In the early 1960s, the application oflogicAnd simulationPsychological activitiesSome commonproblem solving Programs that can prove theorems and performlogical reasoning。However, these general methods cannot solve large practical problems, and it is difficult to transform practical problems into forms suitable for computer solutions, and it is also difficult to deal with the huge search space required for solving problems.E.A. Fegan, 1965BaumOthers are summarizing the generalproblem solving Based on the success and failure experience of the system, combined with the expertise in the field of chemistry, the world's first expert system dendral has been developed, which can infer that chemistrymolecular structure。For more than 20 years,knowledge engineering The theory and technology of expert system are constantly developing, and its application has penetrated into almost every field, including chemistry, mathematics, physics, biology, medicine, agriculture, meteorologygeological prospecting, militaryEngineering technology, legal, commercialspace technology、Auto-Control, computer design and manufacturing, and many of them have reached or even exceeded the level of human experts in the same field in terms of function, and have produced hugeeconomic performance。
The development of expert system has gone through three stages, and is transitioning to the fourth generation.The first generation expert systems (dendal, macsyma, etc.) are characterized by high specialization and strong ability to solve special problems.But inArchitectureIntegrityPortability. There are defects in the transparency and flexibility of the system, and the ability to solve problems is weak.The second generation expert system (Mycin, Casnet, Prospector, Hearsay, etc.) is a single discipline professional and applied system with a complete architecture and improved portabilityman-machine interface , explanation mechanism, knowledge acquisition technology, uncertain reasoning technology, and expert system enhancementknowledge representation andReasoning methodThe enlightenment and versatility of are improved.The third generation expert system is a multidisciplinary integrated system, which adopts a variety ofAI Language, comprehensive use of various knowledge representation methods and multipleInference engineMake andcontrol strategy And began to use all kinds of knowledgeEngineering language, skeleton system andExpert system development toolAnd environment.After summarizing the design methods andimplementation technique On the basis of, it has begun to adopt large-scale multi expert cooperation system, multiple knowledge representation, and integrationknowledge baseSelf organized problem-solving mechanism, multidisciplinary collaborative problem-solving and parallel reasoningExpert system toolsAnd environmentartificial neural networkKnowledge acquisition andLearning mechanismAnd the latest artificial intelligence technology to realize the fourth generation expert system with multiple knowledge bases and multi-agent.
Composition of expert system
Announce
edit
structure
Expert systems are usuallyHuman computer interaction interface, knowledge base, inference engine, interpreter, comprehensive database, knowledge acquisition, etc.In particular, the knowledge base and inference engine are separated from each other and have their own characteristics.The architecture of expert system follows expertsType of system, function and scale.
To enable computers to use expertdomain knowledge , knowledge must be expressed in a certain way.At present, the commonly used knowledge representation methods include production rulessemantic network , Framestate space , logical mode, script, processobject-orientedEtc.Rule-basedProduction systemIt is the most basic method to realize knowledge application at present.The production system consists of three main parts: comprehensive database, knowledge base and inference engine. The comprehensive database contains the world wide facts and assertions for solving problems.The knowledge base contains all knowledge rules expressed in the form of "if: (premise), then: (result)".Inference engine (also called ruleinterpreter)Our task is to use control strategies to find applicable rules.
knowledge base
The knowledge base is used to store the knowledge provided by experts.Expert systemproblem solving The process simulates experts through knowledge in the knowledge baseMode of thinkingTherefore, the knowledge base is the key to whether the quality of the expert system is superior, that is, the quality and quantity of knowledge in the knowledge base determine the quality level of the expert system.Generally speaking, the knowledge base in the expert system is independent of the expert system program. Users can improve the performance of the expert system by changing and improving the knowledge content in the knowledge base.
Knowledge representation forms in artificial intelligence includeProduction, framework, semantic network, etc., and the knowledge that is commonly used in expert systems is production rules.Production rules appear in the form of IF... THEN..., likeBASICetc.programing languageInConditional statementSimilarly, IF is followed by condition (front part), and THEN is followed by conclusion(Afterpart), conditions and conclusions can be passedLogical operationAND, OR, NOT.Here, the production rule is very simple: if the preconditions are met, the corresponding action or conclusion will be generated.
Inference engine
The inference engine repeatedly matches the rules in the knowledge base according to the conditions or known information of the current problem to obtain new conclusionsproblem solving result.Here, the reasoning mode can have positive andBackward reasoningTwo.
The forward chain strategy is to find out the rules whose premises can match the facts or assertions in the database, and use the conflict elimination strategy to select an execution from these rules that can be met, thus changing the content of the original database.In this way, the search is repeated until the fact of the database is consistent with the target, or until there is no rule to match it.
The strategy of the reverse chain is to start from the selected target and find the rules that the execution consequences can reach the target;If the premise of this rule matches the facts in the database, the problem will be solved;Otherwise, take the premise of this rule as a new sub target, find applicable rules for the new sub target, and execute the premise of the reverse sequence until the premise of the last applied rule can match the facts in the database, or until no rules can be applied again, the system will request the user to answer and input the necessary facts in the form of dialogue.
It can be seen that the inference engine is just like the way of thinking of experts to solve problems, and the knowledge base realizes its value through the inference engine.
Other parts
Man machine interface is the interface when the system communicates with users.Through this interface, users can input basic information, answer relevant questions raised by the system, and output reasoning results and relevant explanations.
The comprehensive database is specially used to store theraw data, intermediate results and final conclusions are often used as temporary storage areas.The interpreter can explain the conclusion and solution process according to the user's questions, thus making the expert system more humane.
Knowledge acquisition is the key to the superiority of expert system knowledge base, and it is also an expertsystem designThrough knowledge acquisition, we can expand and modify the content of the knowledge base, and also realize the automatic learning function.
Implementation mode
The early expert system adopted the generalProgramming language(such as fortran, pascal, basic, etc.) and artificial intelligence languages (such as lisp, prologsmalltalkAnd so on), through the cooperation between artificial intelligence experts and domain experts, and through direct programming.Its development cycle is long and difficult, but it is flexible and practical, and is still used by artificial intelligence experts.Most expert systems have been developed using expert systemsdevelopment environment Or expert system development tools. Domain experts can choose appropriate tools to develop their own expert systems, which greatly shortens the development cycle of expert systems, thus providing conditions for the extensive application of expert systems in various fields.
working process
Announce
edit
Structure diagram of expert system
Expert systemBasic structureAs shown in the figure, the arrow direction is the direction of data flow.
The basic workflow of the expert system is that the userinterfaceAnswer questions from the system,Inference engineMatch the information entered by the user with the conditions of each rule in the knowledge base, and store the conclusion of the matched rule inComprehensive databaseMedium.Finally, the expert system will draw the final conclusion and present it to the user.
Here, the expert system can alsointerpreterExplain the following questions to the user: Why does the system ask the user this question?How did the computer come to the final conclusion?
Domain experts orKnowledge EngineerThrough dedicatedsoftware tool, or programming to achieve knowledge acquisition in the expert system, constantly enrich and improve the knowledge in the knowledge base.
(2) Storage detailsproblem solving The initial data of and various information involved in the reasoning process, such as intermediate results, goalsalphabetAnd assumptions.
⑶ According to the current input data, using existing knowledgeReasoning strategy, to solve current problems, and be able to control and coordinate the whole system.
⑷ Be able to make necessary explanations for the reasoning process, conclusions or the system's own behavior, such as problem solving steps, processing strategies, choicesprocessing methodThe reason of the system, the ability of the system to solve certain problems, and how the system organizes and manages its own knowledge.This is not only convenient for users to understand and accept, but also convenient for system maintenance.
(5) Provisionknowledge acquisition, machine learning and knowledge base modification, expansion and improvement.Only in this way can the system'sproblem solving Capability andaccuracy。
(6) Provide oneUser interfaceIt is not only convenient for users to use, but also easy to analyze and understand various requirements and requests of users.
It is emphasized here that storing knowledge and using knowledgeproblem solving It is the two most basic functions of expert system.
characteristic
Expert system is a knowledge-based system, which uses human experts to provideExpertise, simulating human expertsthinking process To solve problems that are quite difficult for human experts.In general, a high-performance expert system should have the following characteristics:
⑴ Inspiration.It can not only use logical knowledge, but also use heuristic knowledge. It uses normative expertise and intuitive judgment knowledge to judge, reason and associateproblem solving 。
(2) Transparency.It enables users toSystem structureIf you don't understand, you can communicate with each other and understand the content of knowledge and reasoning ideas. The system can also answer some questions about the system's own behavior from users.
⑶ Flexibility.The separation of knowledge and reasoning mechanism of expert system enables the system to continuously accept new knowledge, thus ensuring the continuous growth of knowledge in the system to meet the needs of business and research.
classification
Announce
edit
⑴ According to knowledge representation technology, it can be divided into: logic based expert systemRule-based expert systemExpert system based on semantic network and expert system based on framework.
⑵ According to the task type, it can be divided into:
Interpretative: it can be used to analyze symbolic data and explain the practical significance of these data.
Predictive: infer the future evolution results of the object according to the past and present conditions of the object.
Diagnostic type: find the fault and defect of the object according to the input information.
Debugging type: provide the troubleshooting scheme determined by yourself.
Maintenance type: specify and implement the plan to correct certain faults.
Planning type: draw up an action plan according to the given goals.
Design type: form the required scheme and drawings according to the given requirements.
Control type: complete the implementation control task.
Educational type: combination of diagnostic type and debugging type, used for teaching and training.
application
Announce
edit
Application fields of expert system
The original expert system is an application of artificial intelligence, but because of its importance and relevanceapplication systemWith its rapid development, it has become a specific type of information system.The term expert system is defined as "knowledge-based expert system(knowledge-Based expert system)System applicationThe human knowledge stored in the computer can solve problems that usually need experts to solve. It can imitate the reasoning process of human experts when solving specific problems, so it can be used by non experts to improveProblem solvingAt the same time, experts can also regard it as an assistant with professional knowledge.Because onhuman societyIn fact, expert resources are quite scarce. With expert systems, this valuable expert knowledge can be widely used.
Experts in recent yearsSystem technologyIt is gradually mature and widely used in engineering, science, medicine, military, commerce, etc., and has achieved fruitful results, even in someapplication area It also exceeds the intelligence and judgment of human experts.Its functional application fields are summarized as follows:
Interpretations - such as testing the lungs (such as PUFF).
Forecast(Prediction)- If it is predicted that corn loss may be caused by black moth (such as PLAN).
Diagnosis - such as diagnosis of bacterial infection in blood (MYCIN).Another example is diagnosing carsDiesel engineCause of faultCATS system.
Troubleshooting(Fault Isolation) - such as telephone troubleshooting system ACE.
Design - such as specially designed small motor spring andCarbon brushMOTORBRUSHDESIGNER.
Planning - famous auxiliary planningIBMThe layout of the main computer architecture, the reinstallation and rearrangement of the expert system CSS, and the PlanPower expert system for auxiliary property management.
Monitoring - For example, monitoring the YES/MVS of IBM MVS operating system.
Debugging - such as investigating the cause of students' subtraction arithmetic errors.
Repair - such as the expert system SECOFOR for repairing the crude oil storage tank.
Scheduling - such as the expert system ISA for manufacturing and transportation scheduling.Another example is the work shop manufacturing step arrangement system.
Instruction - such as TVC expert system that teaches users to learn operating system.
Control - PTRANS, a control system that helps Digital Corporation manufacture and distribute computers.
Analysis - such as the expert system DIPMETER and analysis of oil well storageOrganic moleculePossible structureDENDRAL system。It is the earliest expert system and also the mostWinnersone of.
Maintenance - such as analysisTelephone exchangeThe expert system COMPASS can recommend how to repair after the failure causes.
Architecture design (Configuration) - such as the expert system XCON for designing VAX computer architecture and the expert system for designing new elevator architectureVTEtc.
Targeting - for example, how to calibrate weapons
example
Here, we take a simple "animal recognition expert system" as an example to preliminarily understand theWorking mechanismAnd system characteristics.The system'sknowledge baseIs aProduction ruleFigure 2 shows two of the rules.In principle, between rulesMutual independenceThe "antecedent" of any two rules cannot be repeated generally, nor does it haveInclusion relationship。The rules of a small expert system can be several or ten, and the rules of a large expert system can be up to thousands. For example, there are only six rules in the knowledge base of this animal expert system.
Trend prediction
At present, at home and abroadExpert system applicationStay in a relatively narrow senseRule-based reasoningAt the basic stage, the application is also more targeted atLaboratory researchAnd someLightweightApplications are far from meeting the requirements of large-scale commercial applications and realizing the requirements of real-time intelligent reasoning and big data processing.
The next step in the development of expert system will be to give priority to model reasoning, supplemented by rule reasoning, meet the needs of commercial applications, and meet the needs of real-time and large data processing.
At the same time, the expert system will develop in a more professional directiondirectionalProvide targeted models and products, such as faults based on the cause and effect directed graph CDGdiagnostic model, ProcessProcessing modelEtc.