CN103761572B - Small expert system reasoning process and conflict set control solution - Google Patents
Small expert system reasoning process and conflict set control solution Download PDFInfo
- Publication number
- CN103761572B CN103761572B CN201410047259.1A CN201410047259A CN103761572B CN 103761572 B CN103761572 B CN 103761572B CN 201410047259 A CN201410047259 A CN 201410047259A CN 103761572 B CN103761572 B CN 103761572B
- Authority
- CN
- China
- Prior art keywords
- rule
- conflict
- expert system
- control solution
- conflict set
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 20
- 230000008569 process Effects 0.000 title claims abstract description 16
- 238000012913 prioritisation Methods 0.000 claims description 4
- 238000013459 approach Methods 0.000 claims description 2
- 230000004069 differentiation Effects 0.000 claims description 2
- 230000009467 reduction Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 201000004569 Blindness Diseases 0.000 description 1
- 230000003213 activating effect Effects 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000004080 punching Methods 0.000 description 1
Abstract
The invention relates to a small expert system reasoning process and conflict set control solution. Overall consideration is carried out mainly from three aspects of reasoning time redundancy, regular structure similarity and rule mode types, forward reasoning is adopted, a conflict set solving strategy based on priority and a chain table is provided, the priority is set according to the rule confidence level, and the higher the confidence level is, the higher the priority is; matching expanses of operation and conflict centralized computing are substantially reduced, and execution efficiency of an expert system inference engine is improved.
Description
Technical field
The present invention relates to specialist system field, especially a kind of small-sized expert system reasoning process and conflict set control are solved
Method.
Background technology
The execution efficiency of productive system is always the focus of Expert System Design personnel.In productive system,
Most of the time is consumed in pattern match work.Therefore study, the pattern matching algorithm of highly effective, to improving system
Efficiency has important meaning.
Propose Rete algorithms within Forgy and 1979 year.This is a kind of Fast Match Algorithm, commercial generation popular at present
Formula system word speed, as OPS, CLIPS etc. are all based on the realization of Rete algorithms.The inference machine of current majority production expert systems
System adopts Rete matching algorithms, or to its improved procedure.Rete algorithms remain conventional matching letter to realize speed
Breath, therefore substantial amounts of memory space is occupied, apply in being more suitable for large-scale specialist system.
Conflict resolution process, refers to after pattern match is carried out, if while have the rule of more than two for election contest rule, being
The process that system has to choose from one to perform.Strategy of Conflict Resolution performs the principle of conflict resolution process institute foundation.
At present, conflict set is designed with following common principle：
(a) nearby principle：The rule of i.e. newest activation is first carried out, and conflict set is designed to a stack first-in last-out.
(b) fairness doctrine：The rule for first activating is first carried out, and conflict set is designed to the stack of first in first out.
(c) priority principle：That is the high rule of priority is first carried out.The evaluation of rule prioritization can by Rules control,
Determine with the mode such as true matching degree.
Nearby principle and the fairness doctrine implement simple, but as no any background information is instructed, can often lead
Reasoning blindness is caused, inference time is slow, the negative consequence such as inefficient；Priority principle efficiency increases than first two,
But a new fact is produced every time, it is necessary to is recalculated priority, is then just dispatched in multiple pending rules,
Increased expense.
The content of the invention
The technical problem to be solved in the present invention is：Propose that a kind of small-sized expert system reasoning process and conflict set control are solved
Method, is considered in terms of inference time redundancy, regular texture similarity and mode of rule type three, is greatly reduced
Matching operation and the expense of conflict centralized calculation, improve the execution efficiency of inference engine of expert system.
The technical solution adopted in the present invention is：A kind of small-sized expert system reasoning process and conflict set control solution party
Method, comprises the following steps：
1) by the rule with model identical structure, set up the annexation of shallow hierarchy；The pattern of strictly all rules is unified
It is numbered, while arranging flag bit for each pattern, initial value is " 0 "；It is that every rule arranges flag bit, initial value is
“0”。
2) in the fact that the initial fact is put into internal memory storehouse；
3) take out unworn true with sentencing that whether the advanced row mode type of all patterns matches from factbase one by one
Not；
4) flag bit for realizing the pattern of matching is configured；
5) by scanning rule storehouse, all patterns are all set to and step 4) in identical rule send into conflict set,
Wait and select to perform；
6) rule to be selected, by the height of credibility, is set up into chained list in conflict set；By rule prioritization with a high credibility
It is just high, it is placed in chained list anterior；By rule with a low credibility, by the rear portion as chained list.
7) node of highest priority in chained list is performed, and new fact is sent in factbase, while the node is deleted；
8) repeat step 3)-step 7), until objective result is exported, or reasoning time-out terminates.
Step 3 of the present invention) in discriminant approach be：True type is different from mode type, then skip, not
Matching；Mode type identical, then carry out matching primitives；By such simple pretreatment, it is to avoid followed by matching primitives
Expense.
Specifically, step 4 of the present invention) in, the flag bit for realizing the pattern of matching is set to into " 1 "；It is described
The step of 5) in, by all patterns by set rule send into conflict set, wait select perform.Described step 7) in, hold
The node of highest priority in row chained list, and new fact is sent in factbase, while the node is deleted, the mark of the rule
Position by set, in representing this reasoning, the rule will not be performed again, it is to avoid redundant operation.
The invention has the beneficial effects as follows：Inference direction of the present invention is forward reasoning, employing based on priority and chained list
Conflict set resolution policy；Priority is set according to Rules control, and the more high then priority of credibility is higher；Greatly reduce
Matching operation and the expense of conflict centralized calculation, improve the execution efficiency of inference engine of expert system.
Description of the drawings
The present invention is further described with reference to the accompanying drawings and examples.
Fig. 1 is rule shallow-layer connection diagram of the invention；
Fig. 2 is reasoning process schematic diagram of the present invention；
Fig. 3 is the basic structure of chained list node of the present invention；
Fig. 4 is the course of work of conflict set of the present invention；
Fig. 5 is execution time reduction rate curve of the present invention.
Specific embodiment
Presently in connection with accompanying drawing and preferred embodiment, the present invention is further detailed explanation.These accompanying drawings are simplified
Schematic diagram, only illustrates the basic structure of the present invention in a schematic way, therefore which only shows the composition relevant with the present invention.
As shown in figure 1, by the rule with model identical structure, the annexation of shallow hierarchy is set up, by strictly all rules
Pattern unification is numbered, and such as i=1,2 ... N. arrange flag bit for each pattern simultaneously, and initial value is " 0 "；It is every rules and regulations
Flag bit is then set, and initial value is " 0 "；
Reasoning detailed process is as shown in Figure 2：
In the fact that the initial fact is put into internal memory the 1st, storehouse；
2nd, the unworn true differentiation with the advanced row mode type of all patterns, true class are taken out from factbase one by one
Type is different from mode type, then skip, and not matches；Mode type identical, then carry out matching primitives；
3rd, the flag bit set of the pattern of matching will be realized；
4th, by scanning rule storehouse, all patterns are sent into into conflict set by the rule of set, is waited and is selected to perform；Punching
It is prominent to concentrate rule to be selected, by the height of credibility, set up chained list.Rule prioritization with a high credibility is just high, is placed in chain
Table is anterior, rule with a low credibility, by the rear portion as chained list.
5th, the node of highest priority in chained list is performed, and new fact is sent in factbase, while the node is deleted,
The regular flag bit is by set；
6th, step 2-5 is repeated, until objective result is exported, or reasoning time-out terminates.
The inference direction be forward reasoning, the conflict set resolution policy based on priority and chained list of employing.Priority is pressed
Set according to Rules control, the more high then priority of credibility is higher.
And it will be seen from figure 1 that when the structural similarity of mode of rule in rule base is higher, then true and pattern match
Number of times will be lower, and in fact, the mode of rule in most of rule base has very high similarity really.By setting up
Such shallow-layer connection, had both avoided the overhead brought by the connections of a large amount of profound levels in Rete algorithms, and preferably real
Show the reduction of matching times, reduce matching expense.
The basic structure of chained list interior joint is as shown in Figure 3.
The course of work of conflict set is as shown in Figure 4.
At present, the inference mechanism of the specialist system much put into commercial operation adopt Rete matching algorithms or its certain in terms of
Improve.In order to verify new algorithm upon execution between on increase, new algorithm is emulated with standard Rete matching algorithm,
Monitoring is tracked to two kinds of algorithms of different run times using JProfiler platforms, run time as shown in Figure 5 is obtained bent
Line.
Abscissa represents regular quantity, and vertical coordinate represents execution time reduction rate.
As can be seen from Figure 5, when regular bar number less than about 400, performed than standard Rete using the execution time of new algorithm
Time is low, and regular number is fewer, and time reduction rate is higher.
Therefore, new algorithm is particularly well-suited to the few small-sized specialist system of fuzzy rules, according to the actual motor fortune set up
The ruuning situation of row fault diagnosis expert system, also demonstrates this point.
The specific embodiment of the simply present invention described in description above, various illustrations reality not to the present invention
Matter Composition of contents is limited, and person of an ordinary skill in the technical field can be in the past described concrete after description has been read
Embodiment is made an amendment or is deformed, without departing from the spirit and scope of invention.
Claims (5)
1. a kind of small-sized expert system reasoning process and conflict set control solution, it is characterised in that comprise the following steps：
1) by the rule with model identical structure, set up the annexation of shallow hierarchy；
2) in the fact that the initial fact is put into internal memory storehouse；
3) take out the unworn true differentiation whether matched with the advanced row mode type of all patterns from factbase one by one；
4) flag bit for realizing the pattern of matching is configured；
5) by scanning rule storehouse, all patterns are all set to and step 4) in identical rule send into conflict set, wait
Select to perform；
6) rule to be selected, by the height of credibility, is set up into chained list in conflict set；
7) node of highest priority in chained list is performed, and new fact is sent in factbase, while the node is deleted；
8) repeat step 3)-step 7), until objective result is exported, or reasoning time-out terminates.
2. a kind of small-sized expert system reasoning process as claimed in claim 1 and conflict set control solution, and its feature exists
In：Described step 1) in, the pattern unification of strictly all rules is numbered, while flag bit is set for each pattern, initially
It is worth for " 0 "；It is that every rule arranges flag bit, initial value is " 0 ".
3. a kind of small-sized expert system reasoning process as claimed in claim 1 and conflict set control solution, and its feature exists
In：Described step 3) in discriminant approach be：True type is different from mode type, then skip, and not matches；Pattern class
Type identical, then carry out matching primitives.
4. a kind of small-sized expert system reasoning process as claimed in claim 1 and conflict set control solution, and its feature exists
In：Described step 4) in, the flag bit for realizing the pattern of matching is set to into " 1 "；Described step 5) in, by all patterns
Conflict set is sent into by the rule of set, is waited and is selected to perform；Described step 7) in, highest priority in execution chained list
Node, and new fact is sent in factbase, while the node is deleted, the regular flag bit is by set.
5. a kind of small-sized expert system reasoning process as claimed in claim 1 and conflict set control solution, and its feature exists
In：Described step 6) in, rule prioritization with a high credibility is just high, it is placed in chained list anterior；By rule with a low credibility,
By the rear portion as chained list.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410047259.1A CN103761572B (en) | 2014-02-11 | 2014-02-11 | Small expert system reasoning process and conflict set control solution |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410047259.1A CN103761572B (en) | 2014-02-11 | 2014-02-11 | Small expert system reasoning process and conflict set control solution |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103761572A CN103761572A (en) | 2014-04-30 |
| CN103761572B true CN103761572B (en) | 2017-04-12 |
Family
ID=50528806
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410047259.1A Expired - Fee Related CN103761572B (en) | 2014-02-11 | 2014-02-11 | Small expert system reasoning process and conflict set control solution |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103761572B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1987913A (en) * | 2005-12-20 | 2007-06-27 | 国际商业机器公司 | Method and system of recommending solutions with an expert system |
| CN101710393A (en) * | 2009-11-25 | 2010-05-19 | 北京航空航天大学 | Method for knowledge expressing and reasoning mechanism of expert system |
| CN103034691A (en) * | 2012-11-30 | 2013-04-10 | 南京航空航天大学 | Method for getting expert system knowledge based on support vector machine |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8898100B2 (en) * | 2011-09-30 | 2014-11-25 | Accenture Global Services Limited | Testing for rule-based systems |
-
2014
- 2014-02-11 CN CN201410047259.1A patent/CN103761572B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1987913A (en) * | 2005-12-20 | 2007-06-27 | 国际商业机器公司 | Method and system of recommending solutions with an expert system |
| CN101710393A (en) * | 2009-11-25 | 2010-05-19 | 北京航空航天大学 | Method for knowledge expressing and reasoning mechanism of expert system |
| CN103034691A (en) * | 2012-11-30 | 2013-04-10 | 南京航空航天大学 | Method for getting expert system knowledge based on support vector machine |
Non-Patent Citations (1)
| Title |
|---|
| 冲突消解策略及其对推理效率的影响;吴鹤龄;《北京理工大学学报》;19951130;第15卷(第5期);13-16 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103761572A (en) | 2014-04-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107943053A (en) | A kind of paths planning method of mobile robot | |
| CN105022377B (en) | A kind of control method of the automated manufacturing system based on Petri network | |
| CN106843211A (en) | A kind of method for planning path for mobile robot based on improved adaptive GA-IAGA | |
| CN109597304A (en) | Die storehouse Intelligent partition storage method based on artificial bee colony algorithm | |
| Liao et al. | A machine learning approach to case adaptation | |
| CN108958202B (en) | Multi-robot collaborative exploration method | |
| CN112828889A (en) | Six-axis cooperative mechanical arm path planning method and system | |
| CN105976401A (en) | Target tracking method and system based on partitioned multi-example learning algorithm | |
| CN111781922A (en) | Multi-robot collaborative navigation method based on deep reinforcement learning and suitable for complex dynamic scene | |
| CN107609694B (en) | Structure optimization method for offshore wind power cluster power transmission system and storage medium | |
| CN103761572B (en) | Small expert system reasoning process and conflict set control solution | |
| Palconit et al. | Diffusion of Robotics in the Philippines: Impact of Policies, Laws, R & D, and Word of Mouth | |
| CN108446814A (en) | With the tree searching method and device of sequential pipeline Job-Shop problem | |
| CN104462853B (en) | Population elite distribution cloud collaboration equilibrium method used for feature extraction of electronic medical record | |
| Petherbridge | 4. Social Solidarity And Intersubjective Recognition: On Axel Honneth’s Struggle For Recognition | |
| CN104615697B (en) | A kind of webpage level remembers return method | |
| CN111275079A (en) | Crowdsourcing label speculation method and system based on graph neural network | |
| CN101853178A (en) | Description method of reconfigurable hardware resource in scheduling | |
| Fujita et al. | Multiagent reinforcement learning with the partly high‐dimensional state space | |
| CN104820621B (en) | Intelligent carriage Synergistic method based on Distributed shared memory | |
| CN105843163B (en) | The circular cutter rail connection method in the section of type containing lake in a kind of island | |
| CN105005842A (en) | Parallel equipment scheduling strategy for just-in-time production | |
| CN102136975A (en) | Large-scale network environment-oriented similarity network construction method | |
| WO2024031831A1 (en) | Mechanical arm packing and unpacking collaboration method and system based on deep reinforcement learning | |
| Audemard et al. | Penelope in sat competition 2014 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170412 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |