CA2037011C - Arrangement for automated troubleshooting using selective advice and a learning knowledge base - Google Patents
Arrangement for automated troubleshooting using selective advice and a learning knowledge baseInfo
- Publication number
- CA2037011C CA2037011C CA002037011A CA2037011A CA2037011C CA 2037011 C CA2037011 C CA 2037011C CA 002037011 A CA002037011 A CA 002037011A CA 2037011 A CA2037011 A CA 2037011A CA 2037011 C CA2037011 C CA 2037011C
- Authority
- CA
- Canada
- Prior art keywords
- user
- signal path
- troubleshooting
- test object
- responsive
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/22—Detection or location of defective computer hardware by testing during standby operation or during idle time, e.g. start-up testing
- G06F11/2257—Detection or location of defective computer hardware by testing during standby operation or during idle time, e.g. start-up testing using expert systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
- G06N5/022—Knowledge engineering; Knowledge acquisition
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y10—TECHNICAL SUBJECTS COVERED BY FORMER USPC
- Y10S—TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y10S706/00—Data processing: artificial intelligence
- Y10S706/902—Application using ai with detail of the ai system
- Y10S706/911—Nonmedical diagnostics
Abstract
AN ARRANGEMENT FOR AUTOMATED TROUBLESHOOTING USING
SELECTIVE ADVICE AND A LEARNING KNOWLEDGE BASE
Abstract Troubleshooting expert systems are generally embodied in software for the purpose of solving difficult problems in some narrow domain of expertise. The prior art describes certain mechanics for developing or generating rules. That process is commonly known as the knowledge acquisition process. Having acquired the knowledge, our new troubleshooting arrangement eliminates the prior art separation between the expert system knowledge acquisition process and the expert system utilization process. Our new arrangement also detects and classifies invalid actions or other errors of the user in a manner that allows for the non-human expert system to advise the human user. Our arrangement interactively communicates between a user and a troubleshooting system, generates a learning knowledge base, identifies an object being tested by the user, utilizes the learning knowledge base for troubleshooting the test object, and classifies the test object as faulty or not faulty. Our arrangement structures prior actions of the user as pairs of elements in a multi-branch tree in the learning knowledge base and the compares present actions of the user with the prior actions of the user. A current signal path is generated between a first input and a first output in the system under test. Then, it can be determined whether the test object belongs to the current signal path. Responsive to the comparison of the actions of the user andresponsive to whether the test object belongs to the current path, the skill level of the user can be classified into one of a plurality of skill levels. Responsive to the skill level classification of the user, the kind of troubleshooting advice to be given the user can be classified as to one of a plurality of kinds of troubleshooting advice and thereafter the advice can be communicated to the user. Further, the output of the test object can be measured and a new current signal path can be generated and re-generated for identifying a faulty object in the system under test.
SELECTIVE ADVICE AND A LEARNING KNOWLEDGE BASE
Abstract Troubleshooting expert systems are generally embodied in software for the purpose of solving difficult problems in some narrow domain of expertise. The prior art describes certain mechanics for developing or generating rules. That process is commonly known as the knowledge acquisition process. Having acquired the knowledge, our new troubleshooting arrangement eliminates the prior art separation between the expert system knowledge acquisition process and the expert system utilization process. Our new arrangement also detects and classifies invalid actions or other errors of the user in a manner that allows for the non-human expert system to advise the human user. Our arrangement interactively communicates between a user and a troubleshooting system, generates a learning knowledge base, identifies an object being tested by the user, utilizes the learning knowledge base for troubleshooting the test object, and classifies the test object as faulty or not faulty. Our arrangement structures prior actions of the user as pairs of elements in a multi-branch tree in the learning knowledge base and the compares present actions of the user with the prior actions of the user. A current signal path is generated between a first input and a first output in the system under test. Then, it can be determined whether the test object belongs to the current signal path. Responsive to the comparison of the actions of the user andresponsive to whether the test object belongs to the current path, the skill level of the user can be classified into one of a plurality of skill levels. Responsive to the skill level classification of the user, the kind of troubleshooting advice to be given the user can be classified as to one of a plurality of kinds of troubleshooting advice and thereafter the advice can be communicated to the user. Further, the output of the test object can be measured and a new current signal path can be generated and re-generated for identifying a faulty object in the system under test.
Description
`~ -1- 2~37011 AN ARRANGEME~T FOR ~UTOMATED TROUBLESIIOOTING USING
SELECTIVE ADVICE AND A LEARNING KNOWLEDGE BASE
Back~round of the Invention Field of the Invention This invention relates to a computer implemented arrangement, which provides effective troubleshooting including fault diagnosis and which uses a hierarchy 5 oE rules with a learning knowledge base.
Description of the Prior Art Expert systems, which are common in the art, are generally embodied in software for the purpose oE solving diEEicult problems in some narrow domain of expertise. Typical examples of expert systems include troubleshooting expert systems, 10 planning or scheduling expert systems and computer aided design expert systems. One example oE a troubleshooting expert system is disclosed in the article by Y. Lirov, entitled "STAREX - simultaneous test and replace circuit pack troubleshooting expert system prototyping and implementation", Engineering Applications of Artificial Intelligence, Vol. 2, No. 1 (March 1989), pp. 3-18. That article discloses an expert 15 system for troubleshooting including diagnosis faults in electronic circuit packs.
Troubleshooting including diagnosing faults typically involves the existence o an error or a fault in some part or parts of a system and involves the use of some procedure to recognize or verify the Eault and to correct the fault.
Troubleshooting expert systems such as the diagnostic expert system 20 disclosed in the Canadian patent application of Y. Lirov and O. Yue, entitled"Technique for Producing an Expert System for System Fault Diagnosis", filed May 31, 1990 as Serial No. 2,017,970 typically include a plurality of modules like a knowledge based module and a sensor based module. The knowledge base module usually includes a set of rules that define conditions and conclusions, which are followed to : 25 solve diagnostic problems, while a typical sensor based module includes a human user, who is equipped with appropriate measurement instruments, called sensors. As an aside and to help understand the concept of rules, the following analog may be considered. The rules in a knowledge base can be compared to a node evaluation function in a typical problem of searching a graph to select the best node for expansion from a current list o candidate nodes whereby the best path to the correct system diagnosis can be found in the shortest amount of time using a minimum of user input. Continuing with our discussion, the rules in the knowledge based module produce directions to a human user in a sensor based module in such a manner that 2 ~
define characteristics such as the "what, when and where" parameters that are to be measured or replaced in the system being troubleshooted. The sensor based module, in turn, communicates the results of the measurements or replacements back to the knowledge base module. In that manner, an interactive selecdon of the most S relevant rules can be accomplished, sometimes through the use of a third module, called an inference engine module. The actual communication in the interaction between the knowledge base module and the sensor based module is usually accomplished through a consultation module. Further details of these and other such modules may be obtained from any standard artificial intelligence textbook. See, for 10 example, P. Winston, "Artificial Intelligence," Academic Press, (New York), 1984.
Unfortunately, it is common that people dislike receiving directions, or advice, for at least three reasons. One reason is that non-human systems tend to be stale unless some process exists to regularly update, replace, change, or alter the knowledge base module of the expert system. In other words it is usually desirable 15 that the non-human system adapt to changes in much the same manner as a humanwould adapt to change. Another reason is that the time spent by the human in receiving the advice and comrnunicating the results of the measurements back to the knowledge base module may be considered to be unproductive, i.e. may be viewed as wasted time. Sdll a third reason is that the non-human system may communicate20 indiscriminate advice instructing the user "what to do" without a satisfactory explanation to the human user as to the reason the user should follow the advice.
The reluctance of human users to advice is particularly noticeable when the advice is communicated from a non-human system.
In our view, the above problems arise in the known art because it is an - 25 accepted procedure to bifurcate certain processes, e.g. it is common to firstly - generate knowledge, or to acquire knowledge, and then to secondly utilize the acquired knowledge in an expert system as two separate and distinct, but repeated, processes. As the underlying process parameters change (e.g., some components in a system may become more likely to fail than others), the expert system may become30 less efficient and therefore it becomes necessary to repeat the knowledge generation ` process of the expert system.
Summary of the Invention This and other problems are solved in accord with the principles of our invention which includes a method, system and apparatus for troubleshooting an 35 object in a system under test. Our arrangement includes interactively communicating between a user and a troubleshooting system, generating a learning knowledge base, 2037~11 identi~ying an object being tested by the user, utilizing the learning knowledge base for trouble shooting the test object, and classifying the test object as faulty or not faulty.
Also, included in our arrangement is a structuring of prior actions of the user as pairs in a multi-branch tree in the learning knowledge base and the comparing of present S actions of the user with the prior actions of the user. A current signal path can be generated between a first input and a first output in the system under test. Then, it can be determined whether the test object belongs to the current signal path.
Responsive to the comparison of the actions of the user and responsive to whether the test object belongs to the current path, the skill level of the user can be classified into 10 one of a plurality of skill levels. Responsive to the sill level classification of the user, `. the kind of troubleshooting advice to be given the user can be classified as to one of a plurality of kinds of troubleshooting advice and thereafter the advice can be communicated to the user. Further, the output of the test object can be measured and a new current signal path can be generated and re-generated for identifying a faulty 15 object in the system under test.
In accordance with one aspect of the invention there is provided a method for troubleshooting an object in a system under test, the method including the steps of interactively communicating between a user and a troubleshooting system, ~, generating a learning knowledge base, identifying an object being tested by the user, 20 utilizing the learning knowledge base for troubleshooting the test object, and classifying the test object as faulty or not faulty AND CHARACTERIZED IN THAT
the method comprises the steps of: (a) structuring prior actions of the user as pairs of elements in a multi-branch tree in the learning knowledge base; (b) responsive to selected pairs of elements, comparing present actions of the user with the prior actions 25 of the user; (c) generating a current signal path between a first input and a first output in the system under test; (d) determining whether the test object belongs to thecurrent signal path; (e) responsive to the comparison of the actions of the user and responsive to whether the test object belongs to the current path, classifying the skill level of the user into one of a plurality of skill levels; (f) responsive to the skill level 30 classification of the user, classifying the kind of troubleshooting advice to be given the user as to one of a plurality of kinds of troubleshooting advice; and (g) communicating the classified troubleshooting advice to the user.
!.,'i :
- 3a -~' In accordance with another aspect of the invention there is provided a system for troubleshooting an object under test, the system including means for interactively communicating between a user and a troubleshooting system, means for - generating a learning knowledge base, means for identifying an object being tested by S the user, means for utilizing the learning knowledge base for troubleshooting the test object, and means for classifying the test object as faulty or not faulty and -- CHARACTERIZED IN THAT the system comprises: means for structuring prior actions of the user as pairs in a multi-branch tree in the learning knowledge base;
means, responsive to selected pairs of elements, for comparing present actions of the user with the prior actions of the user; means for generating a current signal path between a first input and a first output in the system under test; means for determining whether the test object belongs to the current signal path; means responsive to the - comparison of the actions of the user and responsive to whether the test object ,~ belongs to the current path for classifying the skill level of the user into one of a . 15 plurality of skill levels; means responsive to the skill level classification of the user for classifying the kind of troubleshooting advice to be given the user as to one of a - plurality of kinds of troubleshooting advice; and means for communicating the : classified troubleshooting advice to the user.
. In accordance with yet another aspect of the invention there is . 20 provided apparatus for troubleshooting an object in a system under test, the apparatus including means for interactively communicating between a user and troubleshooting . system, means for generating a learning knowledge base, means for identifying an . object being tested by the user, means for utilizing the learning knowledge base for troubleshooting the test object, and means for classifying the test object as faulty or not faulty and CHARACTERIZED IN THAT the apparatus comprises: means for structuring prior actions of the user as pairs in a multi-branch tree in the learning knowledge base; means, responsive to selected pairs of elements, for comparing ~ present actions of the user with the prior actions of the user; means for generating a : current signal path between a first input and a first output in the system under test;
means for determining whether the test object belongs to the current signal path;
means responsive to the comparison of the actions of the user and responsive to whether the test object belongs to the current path for classifying the skill level of the ,~
'.`
,.
,, , ,, .. , . ,-,,, , ~ .
.~.
-3b- 2~37~11 user into one oE a plurality oE skill levels; means responsive to the skill level classiEication oE the user for classiEying the kind oE troubleshooting advice to be given the user as to one of a plurality oE kinds of troubleshooting advice; and means for communicating the classiEied troubleshooting advice to the user.
S The present invention will be described in detail hereinbelow with the aid of the accompanying drawings, in which:
Fig. 1 is a Elow chart used in an explanation of the present invention;
and Fig. 2 is a block diagram for use in explaining the present invention.
The prior art describes certain mechanics for developing or generating rules which can be embodied in a knowledge base module of a troubleshooting expert system. That process is commonly known as the knowledge acquisition process in the artificial intelligence art. Having acquired the knowledge, or having generated the rules, we propose a new troubleshooting arrangement using expert systems, which arrangement eliminates the prior art separation between the expert system knowledge acquisition process and the expert system utilization process. Furthermore, our new arrangement advantageously detects and classifies invalid actions or other errors of the user in a manner that allows for the non-human expert system to advise, or coach, or ` tutor, the human user.
FIG. 1 illustrates the overall information flow using our new arrangement, which includes a method, a system and apparatus and which is based upon six components and which may be suitably embodied in software. We will now briefly describe each of the six components and follow the brief description with a more detailed description of each of the six components as an aid in understanding the principles of our invention.
The first component is a graphic input interface 120, which allows a ; user 110 to communicate specific measurements or replacement actions to the troubleshooting expert system.
~.i .
-` 2~3~
The second component is a learning knowledge base 130 in which is stored any valid action of the user 110. Knowledge base 130 may be embodied as an adjustable weighted multi-branch tree where each of the nodes of the tree corresponds to a condition of a rule and where each of the arcs of the tree 5 corresponds to a conclusion of a rule.
The third component is an user action classifier 140, which may be embodied in a software table. Action classifier 140 allows the expert system to infer and classify any of a plurality of degrees of user troubleshooting expertise by comparing presen~ actions of user 110 as obtained by interface 120 versus prior 10 actions of user 110. The prior actions of user 110 are structured in the multi-branch tree of knowledge base 130. In one illustrative embodiment, action classifier 140 may allow the expertise of user 110 to be classified as one of three levels of troubleshooting skill, for example, (a) as a beginner troubleshooter, or (b) as a good troubleshooter, or (c) as an expert troubleshooter.
The fourth component is an user advice classifier 150, which may also be embodied in a software table. Advice classifier 150 allows the expert system to infer and classify any of a plurality of appropriate kinds of advice responsive to the skill level of user troubleshoo~ng expertise as provided by user acdon classifier 140.
In one illustradve embodiment, advice classifier 150 may allow the advice to user 20 110 to be classified, for exarnple, (a) as troubleshooting strategy advice or (b) as circuit pack schemadc advice.
The fifth component is signal path manager 160, which may also be embodied in software as a set of rules. Signal path rnanager 160 recursively generates and re-generates signal paths (signal path is a term which we will define 25 later) to locate a faulty object.
The sixth component is graphic output interface 170 for communicating ~: the specific advice to usçr 110.
Now we turn to a more detailed descripdon of each of the six components.
The first component, graphic input interface 120, could be embodied (a) as a digidzer, or (b) as a cathode ray tube device with a light pen or other computer input device, or, for that matter, (c) as any device that would permit user 110 to identify objects, which are being measured by user 110, and to cornmunicate the identity of the object to the troubleshooting expert system.
:"
- 203~ 01~
.
s - The second component, learning knowledge base 130, could be embodied as a two-tuple pair of elements, also referred to herein as a pair, including (a) a weight elernent, which reflects some weight or measure of significance to be given the pair, for example, the frequency of measuring an object, and (b) a recursive S data structure element, for example, a tree element. Consider the pair of elements depicted as the tw~tuple ( weight, tree ) where the tree element is defined using the following recursive symbology:
tree ( object, advice, if-fail ( list of pairs ), if-pass ( list of pairs) ) (1)The symbology used in equadon (1) is a short hand notation for a recursive 10 data structure that includes four data fields, which we now describe.
The first data field, called the object field, includes an encoded identification of the object being measured by user 110. For example, assume that the object being measured is an intergrated circuit. It may be desirable to measure specific components or other parts of the intergrated circuit, which themselves would be objects for the purposes 15 of illustradng the principles of our invention. With that background, the object field in equadon (1) could be encoded to, for example, identify a specific pin on a specific intergrated circuit, which itself could be a component of a circuit pack.
The second data field, called the advice field, includes an encoded statement of advice to be communicated to user 110. As one example, the statement of advice 20 could recornmend a procedure, which is to be followed so as to measure the object idendfied in the object field. As another example, the advice field could be a statement to the effect that the object should be tested with a voltmeter for some period of time and that the results of the voltmeter test should be stored for subsequent processing. The advice field can be formulated in any of a number of ways. One way to formulate the 25 advice field is to define it in such a manner that it expects a binary result relative to the statement of advice. In that manner, the outcome of the test that is recommended in the - advice field can be characterized as either having failed (in which case our method either uses a pair of elements from the list of pairs in the third data field or creates a new pair for insertion in the third data field) or being passed (in which case our method either uses 30 a pair from the list of pair in the fourth data field or creates a new pair for insertion in the fourth field). This leads to a more detailed description of the last two fields in equadon : (1).
The third data field, called the if-fail ( list of pairs ) field, includes a list of ~' empty or one or more pairs, each of which includes both a weight element and a tree ~- 35 element, which may be defined using equation (1).
' ., i ., 203~
The fourth data field, called the if-pass ( list of pairs ) field, includes, like the third data field, a list of empty or one or more pairs, which include both a weight element and a tree element, also defined pursuant to equadon (1).
As an ~; ~ ;e, if both the third and fourth data fields include an empty list of5 pairs, there is no further recursion beyond the object idendfied in the first data field of that tree element. Such a tree with an empty list of successor pairs is also called a leaf in the artificial intelligence art. An empty list of pairs is simply a set of pairs which has no elements.
We now turn to a more specific descripdon of the contents of the third and 10 fourth data fields and of the meaning of the phrase list of pairs.
Each list of pairs in the third and fourth data fields is a rank ordering of oneor more pairs where the rank ordering is in one-to-one correspondence with the relative magnitude of the weight element of a pair. Also, the next object to be measured will be that object encoded in the object field of the tree element in the first pair in the list of 15 pairs. That result occurs in view of the rank ordering of pairs in each list of pairs. That is, the first pair in a list of pairs will have a weight element, which is generally larger and in any case is not less than the weight element of any other pair in the list of pairs.
Either the third field or the fourth field of a tree element is selected in response to the binary result obtained from applying the advice stated in the second field 20 on the object stated in the first field of the tree element of a two-tuple pair of elements.
That is, the third field is selected if the advice, when applied, fails, while the fourth field is selected if the advice, when applied, passes.
` The third component, user action classifier 140, could be embodied as a set - of rules that characterize or classify the skill, or the skill level, of user 110 as one skill 25 level out of a plurality of possible troubleshooting skill levels. For example, the rules could be established in a manner that classify the user's skill (a) as a beginner troubleshooter, or (b) as a good troubleshooter, or (c) as an expert troubleshooter. With ~' such a three skill level assumpdon and with the added assumpdon that the troubleshoodng expert system includes a first memory that maintains the identification of 30 a current list of pairs of the type included in learning knowledge base 130, the following -' set of rules can be embodied in user acdon classifier 140:
Rule 14~1: If the test object selected by user 110 is the same object as the object idendfied in the first data field of the tree element of one of dle pairs in the current list of pairs, hereinafter that one pair is called the present pair, then the weight element of 35 the present pair is incremented, in our illustrative embodiment, by a numerical valuç of one. The current list of pairs with the incremented present pair is then sorted, which, 2~3'~0 11 because of the rank ordering by weight, may result in the present pair being relocated to a different relative position within the current list of pairs. In addition, (a) the skill of user 110 is classified as a good troubleshooter, and (b) the present pair from the current list of pairs is moved, through use of its identification in the first memory of the troubleshooting 5 expert system, to a second memory of the troubleshooting expert system and becomes known as the previous pair, i.e. the previous pair is stored in the second memory, and (c) a new list of pairs is selected and the identification of the new list replaces the identification of the current list of pairs in the first memory of the trouUeshooting expert system, and (d) the new list of pairs replaces the current list of pairs as the new current 10 list of pairs.
Ripe for discussion at this juncture is the process by which the new list of pairs is selected. As previously described, the third data field and the fourth data field of a tree element includes a list of pairs. As also previously described, the third data field is selected upon detecting of a failure of the advice offered in the advice field and the fourth 15 data field is selected upon detecting of a passing of the advice offered in the advice field of the tree element in the present pair. Accordingly, in response to the binary outcome of the test either having failed or being passed, respectively, a new list of pairs is selected.
Rule 140-2: If the test object selected by user 110 is not the same object as the object idendfied in the first data field of a tree element of any of the pairs in the 20 current list of pairs, then the skill level of user 110 is classified as not a good troubleshooter. Not a good troubleshooter in our illustradve embodiment means that the skill level of user 110 is either that of a beginner troubleshooter or that of an expert troubleshooter.
As an aside, it can be mentioned that, in view of Rule 140-2 above, the 25 troubleshooting expert system begins to adapt its learning knowledge base 130. That is, the non-human troubleshooting expert system adapts to changes in much the same manner as a human would adapt to change.
Condnuing, in response to the "If ' condidon of Rule 140-2, user action classifier 140 can be suitably programrned to perform the following methodology, which 30 includes generadng a new two-tuple pair with a weight element set to a numerical value of one and with a tree element generated in a manner which we will describe in the next paragraph. Upon generadng the new pair, ~he present pair is moved to the second memory of the trouUeshoodng expert system and becomes known as the previous paitwhile the newly created, or generated, pair is inserted in its Tank ordered location in the 35 curtent list of pairs, which is idendfied ~n the first memory of the troubleshooting expert system, and becomes known as the nt pair. In that mannet the troubleshoodng - 203 ~
:~
expert system adapts its learning knowledge base 130. Noteworthy, the system adapts concurrently with the use of the system, and not, as in the prior art, as separate and distinct processes.
- We now descri'oe a methodology for generating the tree element of the newly S generated pair.
The first data, or object, field of the tree is set to an encoded identification of the object selected by user 110.
The second data, or advice, field of the tree, which is an encoded statement of procedure that is recommended to be followed when measuring the object identified in 10 the object field, is generated in response to an interaction between user 110 and the troubleshooting expe~t system by way of which user 110 specifies the procedure to measure the object selected by user 110.
The third data, or if-fail, field as well as the fourth data, or if-pass, field are encoded to include empty lists of two-tuple pairs. Inasmuch as both the third and the 15 fourth data fields of the newly generated tree element include empty lists of pairs, that tree is called a leaf.
We now turn to FIG. 2 as an aid in defining a concept called signal path, which will be used in a further descripdon of user acdon classifier 140, user advice classifier 150, and signal path manager 160.
FIG. 2 illustrates a unit 200, which is a generic object, that includes a plurality of objects like objects 221, 224, 229 among others, which are interconnected in some predefined order to provide some funcdon. For purposes of a descripdon of an illustradve embodiment of the principles of our invendon, the particular function is not important and therefore its descripdon is omitted. We assume that a user like user 110 .~: 25 desires to troubleshoot unit 200 and does so in accord with a certain set of rules. Each object as well as the unit itself has an input and an output. For example, input 201 and output 232 are the input and output respectively of unit 200. Input 201 is also the input to object 221 while output 232 is also the output of object 231. Using that background, we now define the concept of a signal path. A signal path is an ordered set of objects 30 from a specific input to a specific output. For example, the signal path from input 241 to output 261 of object 260 includes object 2æ with its output 251, which is also the input 251 to object 250. The output 261 of object 250 is also the output 261 of object 260.
Object 250 has a signal path from its input 251 to its output 261 through both of objects 224 and 225.
g Now the concept of a signal path can be further discussed in terms of a downstream signal path and an upstream signal path.
As mentioned, a signal path is an ordered set of objects from a specific input to a specific output. On the one hand, for a first signal path having an object in the first S signal path, the ~erm downstream signal path means a ~econd signal path consisting of that part of the first signal path, which extends from the output of the object in question and continues to the output of the first signal path. On the other hand, for a first signal path having an object in the first sigDal path, the term upstream signal path means a second signal path consisting of that part of the first signal path, which extends from the 10 input of the first signal path and continues to the input of the object in question. Note that the object in question is itself excluded from the downstream signal path and is also excluded from the upstream signal path. For example, consider a signal path defined from input 241 through objects 260 and 231 to output 232. Now consider object 250. Relative to object 250, the downstream signal path includes object 231 while the upstream signal 15 path includes object 222. As another example, consider a signal path defined from input 201 through object 200 to output 232. Again consider object 250. Relative to object 250, the downstream signal path includes object 231 as before; however, the upstream signal path includes objects 221 and 222. Note that object 270 as well as the plurality of objects illustrated in object 270 are excluded from both the downstream and upstream - ~ 20 signal paths when defined relative to object 250 even though object 270 as well as objects -~ 223, 226, 227, 228, 229 and 230 within object 270 are in the signal path from input 201 through object 200 to output 232.
Now the concept of a signal path can be further discussed in terms of a current signal path.
Consider any signal path with the property that its input is found to be correct while its output is found to be in fault. We can conclude that there is some malfunction in the signal path. The term current signal path means a signal path in which a fault is detected. An encoded identity of the current signal path is stored in a third - memory of the troubleshooting expert system.
It is now convenient to offer a further definition of troubleshooting.
; Troubleshooting can be viewed as the process of idendfying a current signal path. With knowledge of the culTent signal path, it becomes the further task of the troubleshooter to locate the faulty object(s) within the current signal path. The troubleshooter can select an object in the current signal path, measure the output of the selected object, and determine 35 whether or not the output is in fault. On the one hand, if the output is faulty, then either the selected object is at fault or an object in an upstrearn signal path is at fault. On the --~ 2~3~01~
- other hand, if the output of the selected object is not at fault, then an object in the downstream signal path is at fault.
We now return to user action classifier 140 to complete its descripdon. One further rule is em~ in action classifier 140:
Rule 140-3: If the test object selected by user 110 does not belong to the current signal path, then the skill level of user 110 is characterized as that of a beginner troubleshooter.
- The object field and the advice field of the tree element in one of the pairs in the current list of pairs is used: (a) for aiding user 110 in selecting the object to be tested 10 from the current signal path, and (b) for comparing the object selected by user 110 with ' the object identified in the object field of the tree element in the first two-tuple pair in the ' current list of pairs, and (c) for inferring and classifying the action of user 110 in response to the comparison results. For example, the user's actions may be classified according to whether user 110 (i) is or is not using a correct strategy and/or (ii) 15 understands or does not understand the causal dependency between the objects in the current signal path.
That completes our descripdon of user action classifier 140.
The fourth component, user advice classifier 150, could be embodied in software as a set of three rules allowing the troubleshooting expert system to infer and ;~ 20 classify the appropriate kind of advice, which ought to be given user 110, in response to the skill level of user 110.
.
Rule 150-1: If the sldll level of user 110 is that of a beginner ~oubleshooter, then it could be concluded that user 110 needs advice that relates to troubleshooting strategy and, accordingly, advice classifier 150 could so set a flag for later use in 25 signalling the need for that strategy to signal path manager 160.
: Rule 150-2: If the skill level of user 110 is that of a beginner troubleshooter or an expert trouUeshooter, then it could be concluded that user 110 needs advice that relates to the current signal path and, accordingly, advice classifier 150 could so set a fiag for later use in signalling the need for that strategy to signal path manager 160.
; 30 Rule 150-3: If the skill level of user 110 is that of a good troubleshooter, then it would be concluded that user 110 does not need advice and, accordingly, advice ~; classifier 150 could so set a flag for later use in signalling the need for that strategy to signal path manager 160.
That completes our description of user advice classifier 150.
- 203~n~ 1 The fifth component, signal path manager 160, could be embodied in software as a set of two rules allowing the expert system to recursively generate and re-generate the current signal path for the pulpose of seeking out the object in the current signal path, which is at fault.
S Rule 160-1: If the output of the object selected by user 110 is not at fault, then a new culrent signa1 path is generated by the process of (a) retrieving the current signal path from the third memory of the troubleshooting expert system; (b) generating an upstream signal path responsive to the selected object and responsive to the culrent signal path; (c) deleting, or erasing, or removing the selected object and all objects of the 10 upstream signal path from the current signal path to obtain a new current signal path; and (d) storing the new current signal path in the third memory.
Rule 160-2: If the output of the object selected by user 110 is at fault, then anew current signal path is generated by the process of (a) retrieving the cuTrent signal path from the third memory of the troubleshooting expert system; (b) generating a 15 downstream signal path responsive to the selected object and responsive to the current signal path; (c) deleting, or erasing, or removing the selected all objects of the f~, downstream signal path from the current signal path to obtain a new cusrent signal path;
and (d) storing the new current signal path in the third memory.
That completes our description of signal path manager 160.
; 20 The sixth component, graphic output interface 170, could be embodied through a set of two rules allowing the expert system to communicate the appropriate advice to user 110 responsive to the kind of advice identified by the flag set by user advice classifier 150 and responsive to the current list of pairs of the type included in learning knowledge base 130 and responsive to the current signal path.
` 25 Rule 17~1: If the flag set by user advice classifier 150 is set to indicate that . user 110 needs advice that relates to the current signal path, then communicate the current signal path to user 110.
Rule 170-2: If the flag set by user advice classifier 150 is set to indicate that - user 110 needs advice that relates to troubleshooting strategy, then communicate the 30 object field and the advise field from the tree element of the first two-tuple pair in the current list of pairs to user 110.
: Note that the first two-tuple pair includes the tree element with the weight element having the greatest value among the current list of pairs and, according to Rule 140-1, that is the tree which will be used most often and therefore is in a sense the best 35 troubleshoodng advice.
~,~
2~3 ~ 011 ;; Note also that pursuant to Rules 170-1 and 170-2, the only advice - ~ communicated to user 110 is advice, which is relevant to the kind of troubleshooting error actually detected by user action classifier 140. Moreover, as user 110 is not required , to follow the advice .lt each and every step of the troubleshooting process, there is less S time wasted by the consultation module, i.e. unproductive time is remarkably decreased.
Although our invention has been described and illustrated in detail, it is to beunderstood that the same is not by way of limitation. Hence, the spirit and scope of our invention is limited only by the terms of the appended claims.
.
., .
.
',' ' .
:.
, .
SELECTIVE ADVICE AND A LEARNING KNOWLEDGE BASE
Back~round of the Invention Field of the Invention This invention relates to a computer implemented arrangement, which provides effective troubleshooting including fault diagnosis and which uses a hierarchy 5 oE rules with a learning knowledge base.
Description of the Prior Art Expert systems, which are common in the art, are generally embodied in software for the purpose oE solving diEEicult problems in some narrow domain of expertise. Typical examples of expert systems include troubleshooting expert systems, 10 planning or scheduling expert systems and computer aided design expert systems. One example oE a troubleshooting expert system is disclosed in the article by Y. Lirov, entitled "STAREX - simultaneous test and replace circuit pack troubleshooting expert system prototyping and implementation", Engineering Applications of Artificial Intelligence, Vol. 2, No. 1 (March 1989), pp. 3-18. That article discloses an expert 15 system for troubleshooting including diagnosis faults in electronic circuit packs.
Troubleshooting including diagnosing faults typically involves the existence o an error or a fault in some part or parts of a system and involves the use of some procedure to recognize or verify the Eault and to correct the fault.
Troubleshooting expert systems such as the diagnostic expert system 20 disclosed in the Canadian patent application of Y. Lirov and O. Yue, entitled"Technique for Producing an Expert System for System Fault Diagnosis", filed May 31, 1990 as Serial No. 2,017,970 typically include a plurality of modules like a knowledge based module and a sensor based module. The knowledge base module usually includes a set of rules that define conditions and conclusions, which are followed to : 25 solve diagnostic problems, while a typical sensor based module includes a human user, who is equipped with appropriate measurement instruments, called sensors. As an aside and to help understand the concept of rules, the following analog may be considered. The rules in a knowledge base can be compared to a node evaluation function in a typical problem of searching a graph to select the best node for expansion from a current list o candidate nodes whereby the best path to the correct system diagnosis can be found in the shortest amount of time using a minimum of user input. Continuing with our discussion, the rules in the knowledge based module produce directions to a human user in a sensor based module in such a manner that 2 ~
define characteristics such as the "what, when and where" parameters that are to be measured or replaced in the system being troubleshooted. The sensor based module, in turn, communicates the results of the measurements or replacements back to the knowledge base module. In that manner, an interactive selecdon of the most S relevant rules can be accomplished, sometimes through the use of a third module, called an inference engine module. The actual communication in the interaction between the knowledge base module and the sensor based module is usually accomplished through a consultation module. Further details of these and other such modules may be obtained from any standard artificial intelligence textbook. See, for 10 example, P. Winston, "Artificial Intelligence," Academic Press, (New York), 1984.
Unfortunately, it is common that people dislike receiving directions, or advice, for at least three reasons. One reason is that non-human systems tend to be stale unless some process exists to regularly update, replace, change, or alter the knowledge base module of the expert system. In other words it is usually desirable 15 that the non-human system adapt to changes in much the same manner as a humanwould adapt to change. Another reason is that the time spent by the human in receiving the advice and comrnunicating the results of the measurements back to the knowledge base module may be considered to be unproductive, i.e. may be viewed as wasted time. Sdll a third reason is that the non-human system may communicate20 indiscriminate advice instructing the user "what to do" without a satisfactory explanation to the human user as to the reason the user should follow the advice.
The reluctance of human users to advice is particularly noticeable when the advice is communicated from a non-human system.
In our view, the above problems arise in the known art because it is an - 25 accepted procedure to bifurcate certain processes, e.g. it is common to firstly - generate knowledge, or to acquire knowledge, and then to secondly utilize the acquired knowledge in an expert system as two separate and distinct, but repeated, processes. As the underlying process parameters change (e.g., some components in a system may become more likely to fail than others), the expert system may become30 less efficient and therefore it becomes necessary to repeat the knowledge generation ` process of the expert system.
Summary of the Invention This and other problems are solved in accord with the principles of our invention which includes a method, system and apparatus for troubleshooting an 35 object in a system under test. Our arrangement includes interactively communicating between a user and a troubleshooting system, generating a learning knowledge base, 2037~11 identi~ying an object being tested by the user, utilizing the learning knowledge base for trouble shooting the test object, and classifying the test object as faulty or not faulty.
Also, included in our arrangement is a structuring of prior actions of the user as pairs in a multi-branch tree in the learning knowledge base and the comparing of present S actions of the user with the prior actions of the user. A current signal path can be generated between a first input and a first output in the system under test. Then, it can be determined whether the test object belongs to the current signal path.
Responsive to the comparison of the actions of the user and responsive to whether the test object belongs to the current path, the skill level of the user can be classified into 10 one of a plurality of skill levels. Responsive to the sill level classification of the user, `. the kind of troubleshooting advice to be given the user can be classified as to one of a plurality of kinds of troubleshooting advice and thereafter the advice can be communicated to the user. Further, the output of the test object can be measured and a new current signal path can be generated and re-generated for identifying a faulty 15 object in the system under test.
In accordance with one aspect of the invention there is provided a method for troubleshooting an object in a system under test, the method including the steps of interactively communicating between a user and a troubleshooting system, ~, generating a learning knowledge base, identifying an object being tested by the user, 20 utilizing the learning knowledge base for troubleshooting the test object, and classifying the test object as faulty or not faulty AND CHARACTERIZED IN THAT
the method comprises the steps of: (a) structuring prior actions of the user as pairs of elements in a multi-branch tree in the learning knowledge base; (b) responsive to selected pairs of elements, comparing present actions of the user with the prior actions 25 of the user; (c) generating a current signal path between a first input and a first output in the system under test; (d) determining whether the test object belongs to thecurrent signal path; (e) responsive to the comparison of the actions of the user and responsive to whether the test object belongs to the current path, classifying the skill level of the user into one of a plurality of skill levels; (f) responsive to the skill level 30 classification of the user, classifying the kind of troubleshooting advice to be given the user as to one of a plurality of kinds of troubleshooting advice; and (g) communicating the classified troubleshooting advice to the user.
!.,'i :
- 3a -~' In accordance with another aspect of the invention there is provided a system for troubleshooting an object under test, the system including means for interactively communicating between a user and a troubleshooting system, means for - generating a learning knowledge base, means for identifying an object being tested by S the user, means for utilizing the learning knowledge base for troubleshooting the test object, and means for classifying the test object as faulty or not faulty and -- CHARACTERIZED IN THAT the system comprises: means for structuring prior actions of the user as pairs in a multi-branch tree in the learning knowledge base;
means, responsive to selected pairs of elements, for comparing present actions of the user with the prior actions of the user; means for generating a current signal path between a first input and a first output in the system under test; means for determining whether the test object belongs to the current signal path; means responsive to the - comparison of the actions of the user and responsive to whether the test object ,~ belongs to the current path for classifying the skill level of the user into one of a . 15 plurality of skill levels; means responsive to the skill level classification of the user for classifying the kind of troubleshooting advice to be given the user as to one of a - plurality of kinds of troubleshooting advice; and means for communicating the : classified troubleshooting advice to the user.
. In accordance with yet another aspect of the invention there is . 20 provided apparatus for troubleshooting an object in a system under test, the apparatus including means for interactively communicating between a user and troubleshooting . system, means for generating a learning knowledge base, means for identifying an . object being tested by the user, means for utilizing the learning knowledge base for troubleshooting the test object, and means for classifying the test object as faulty or not faulty and CHARACTERIZED IN THAT the apparatus comprises: means for structuring prior actions of the user as pairs in a multi-branch tree in the learning knowledge base; means, responsive to selected pairs of elements, for comparing ~ present actions of the user with the prior actions of the user; means for generating a : current signal path between a first input and a first output in the system under test;
means for determining whether the test object belongs to the current signal path;
means responsive to the comparison of the actions of the user and responsive to whether the test object belongs to the current path for classifying the skill level of the ,~
'.`
,.
,, , ,, .. , . ,-,,, , ~ .
.~.
-3b- 2~37~11 user into one oE a plurality oE skill levels; means responsive to the skill level classiEication oE the user for classiEying the kind oE troubleshooting advice to be given the user as to one of a plurality oE kinds of troubleshooting advice; and means for communicating the classiEied troubleshooting advice to the user.
S The present invention will be described in detail hereinbelow with the aid of the accompanying drawings, in which:
Fig. 1 is a Elow chart used in an explanation of the present invention;
and Fig. 2 is a block diagram for use in explaining the present invention.
The prior art describes certain mechanics for developing or generating rules which can be embodied in a knowledge base module of a troubleshooting expert system. That process is commonly known as the knowledge acquisition process in the artificial intelligence art. Having acquired the knowledge, or having generated the rules, we propose a new troubleshooting arrangement using expert systems, which arrangement eliminates the prior art separation between the expert system knowledge acquisition process and the expert system utilization process. Furthermore, our new arrangement advantageously detects and classifies invalid actions or other errors of the user in a manner that allows for the non-human expert system to advise, or coach, or ` tutor, the human user.
FIG. 1 illustrates the overall information flow using our new arrangement, which includes a method, a system and apparatus and which is based upon six components and which may be suitably embodied in software. We will now briefly describe each of the six components and follow the brief description with a more detailed description of each of the six components as an aid in understanding the principles of our invention.
The first component is a graphic input interface 120, which allows a ; user 110 to communicate specific measurements or replacement actions to the troubleshooting expert system.
~.i .
-` 2~3~
The second component is a learning knowledge base 130 in which is stored any valid action of the user 110. Knowledge base 130 may be embodied as an adjustable weighted multi-branch tree where each of the nodes of the tree corresponds to a condition of a rule and where each of the arcs of the tree 5 corresponds to a conclusion of a rule.
The third component is an user action classifier 140, which may be embodied in a software table. Action classifier 140 allows the expert system to infer and classify any of a plurality of degrees of user troubleshooting expertise by comparing presen~ actions of user 110 as obtained by interface 120 versus prior 10 actions of user 110. The prior actions of user 110 are structured in the multi-branch tree of knowledge base 130. In one illustrative embodiment, action classifier 140 may allow the expertise of user 110 to be classified as one of three levels of troubleshooting skill, for example, (a) as a beginner troubleshooter, or (b) as a good troubleshooter, or (c) as an expert troubleshooter.
The fourth component is an user advice classifier 150, which may also be embodied in a software table. Advice classifier 150 allows the expert system to infer and classify any of a plurality of appropriate kinds of advice responsive to the skill level of user troubleshoo~ng expertise as provided by user acdon classifier 140.
In one illustradve embodiment, advice classifier 150 may allow the advice to user 20 110 to be classified, for exarnple, (a) as troubleshooting strategy advice or (b) as circuit pack schemadc advice.
The fifth component is signal path manager 160, which may also be embodied in software as a set of rules. Signal path rnanager 160 recursively generates and re-generates signal paths (signal path is a term which we will define 25 later) to locate a faulty object.
The sixth component is graphic output interface 170 for communicating ~: the specific advice to usçr 110.
Now we turn to a more detailed descripdon of each of the six components.
The first component, graphic input interface 120, could be embodied (a) as a digidzer, or (b) as a cathode ray tube device with a light pen or other computer input device, or, for that matter, (c) as any device that would permit user 110 to identify objects, which are being measured by user 110, and to cornmunicate the identity of the object to the troubleshooting expert system.
:"
- 203~ 01~
.
s - The second component, learning knowledge base 130, could be embodied as a two-tuple pair of elements, also referred to herein as a pair, including (a) a weight elernent, which reflects some weight or measure of significance to be given the pair, for example, the frequency of measuring an object, and (b) a recursive S data structure element, for example, a tree element. Consider the pair of elements depicted as the tw~tuple ( weight, tree ) where the tree element is defined using the following recursive symbology:
tree ( object, advice, if-fail ( list of pairs ), if-pass ( list of pairs) ) (1)The symbology used in equadon (1) is a short hand notation for a recursive 10 data structure that includes four data fields, which we now describe.
The first data field, called the object field, includes an encoded identification of the object being measured by user 110. For example, assume that the object being measured is an intergrated circuit. It may be desirable to measure specific components or other parts of the intergrated circuit, which themselves would be objects for the purposes 15 of illustradng the principles of our invention. With that background, the object field in equadon (1) could be encoded to, for example, identify a specific pin on a specific intergrated circuit, which itself could be a component of a circuit pack.
The second data field, called the advice field, includes an encoded statement of advice to be communicated to user 110. As one example, the statement of advice 20 could recornmend a procedure, which is to be followed so as to measure the object idendfied in the object field. As another example, the advice field could be a statement to the effect that the object should be tested with a voltmeter for some period of time and that the results of the voltmeter test should be stored for subsequent processing. The advice field can be formulated in any of a number of ways. One way to formulate the 25 advice field is to define it in such a manner that it expects a binary result relative to the statement of advice. In that manner, the outcome of the test that is recommended in the - advice field can be characterized as either having failed (in which case our method either uses a pair of elements from the list of pairs in the third data field or creates a new pair for insertion in the third data field) or being passed (in which case our method either uses 30 a pair from the list of pair in the fourth data field or creates a new pair for insertion in the fourth field). This leads to a more detailed description of the last two fields in equadon : (1).
The third data field, called the if-fail ( list of pairs ) field, includes a list of ~' empty or one or more pairs, each of which includes both a weight element and a tree ~- 35 element, which may be defined using equation (1).
' ., i ., 203~
The fourth data field, called the if-pass ( list of pairs ) field, includes, like the third data field, a list of empty or one or more pairs, which include both a weight element and a tree element, also defined pursuant to equadon (1).
As an ~; ~ ;e, if both the third and fourth data fields include an empty list of5 pairs, there is no further recursion beyond the object idendfied in the first data field of that tree element. Such a tree with an empty list of successor pairs is also called a leaf in the artificial intelligence art. An empty list of pairs is simply a set of pairs which has no elements.
We now turn to a more specific descripdon of the contents of the third and 10 fourth data fields and of the meaning of the phrase list of pairs.
Each list of pairs in the third and fourth data fields is a rank ordering of oneor more pairs where the rank ordering is in one-to-one correspondence with the relative magnitude of the weight element of a pair. Also, the next object to be measured will be that object encoded in the object field of the tree element in the first pair in the list of 15 pairs. That result occurs in view of the rank ordering of pairs in each list of pairs. That is, the first pair in a list of pairs will have a weight element, which is generally larger and in any case is not less than the weight element of any other pair in the list of pairs.
Either the third field or the fourth field of a tree element is selected in response to the binary result obtained from applying the advice stated in the second field 20 on the object stated in the first field of the tree element of a two-tuple pair of elements.
That is, the third field is selected if the advice, when applied, fails, while the fourth field is selected if the advice, when applied, passes.
` The third component, user action classifier 140, could be embodied as a set - of rules that characterize or classify the skill, or the skill level, of user 110 as one skill 25 level out of a plurality of possible troubleshooting skill levels. For example, the rules could be established in a manner that classify the user's skill (a) as a beginner troubleshooter, or (b) as a good troubleshooter, or (c) as an expert troubleshooter. With ~' such a three skill level assumpdon and with the added assumpdon that the troubleshoodng expert system includes a first memory that maintains the identification of 30 a current list of pairs of the type included in learning knowledge base 130, the following -' set of rules can be embodied in user acdon classifier 140:
Rule 14~1: If the test object selected by user 110 is the same object as the object idendfied in the first data field of the tree element of one of dle pairs in the current list of pairs, hereinafter that one pair is called the present pair, then the weight element of 35 the present pair is incremented, in our illustrative embodiment, by a numerical valuç of one. The current list of pairs with the incremented present pair is then sorted, which, 2~3'~0 11 because of the rank ordering by weight, may result in the present pair being relocated to a different relative position within the current list of pairs. In addition, (a) the skill of user 110 is classified as a good troubleshooter, and (b) the present pair from the current list of pairs is moved, through use of its identification in the first memory of the troubleshooting 5 expert system, to a second memory of the troubleshooting expert system and becomes known as the previous pair, i.e. the previous pair is stored in the second memory, and (c) a new list of pairs is selected and the identification of the new list replaces the identification of the current list of pairs in the first memory of the trouUeshooting expert system, and (d) the new list of pairs replaces the current list of pairs as the new current 10 list of pairs.
Ripe for discussion at this juncture is the process by which the new list of pairs is selected. As previously described, the third data field and the fourth data field of a tree element includes a list of pairs. As also previously described, the third data field is selected upon detecting of a failure of the advice offered in the advice field and the fourth 15 data field is selected upon detecting of a passing of the advice offered in the advice field of the tree element in the present pair. Accordingly, in response to the binary outcome of the test either having failed or being passed, respectively, a new list of pairs is selected.
Rule 140-2: If the test object selected by user 110 is not the same object as the object idendfied in the first data field of a tree element of any of the pairs in the 20 current list of pairs, then the skill level of user 110 is classified as not a good troubleshooter. Not a good troubleshooter in our illustradve embodiment means that the skill level of user 110 is either that of a beginner troubleshooter or that of an expert troubleshooter.
As an aside, it can be mentioned that, in view of Rule 140-2 above, the 25 troubleshooting expert system begins to adapt its learning knowledge base 130. That is, the non-human troubleshooting expert system adapts to changes in much the same manner as a human would adapt to change.
Condnuing, in response to the "If ' condidon of Rule 140-2, user action classifier 140 can be suitably programrned to perform the following methodology, which 30 includes generadng a new two-tuple pair with a weight element set to a numerical value of one and with a tree element generated in a manner which we will describe in the next paragraph. Upon generadng the new pair, ~he present pair is moved to the second memory of the trouUeshoodng expert system and becomes known as the previous paitwhile the newly created, or generated, pair is inserted in its Tank ordered location in the 35 curtent list of pairs, which is idendfied ~n the first memory of the troubleshooting expert system, and becomes known as the nt pair. In that mannet the troubleshoodng - 203 ~
:~
expert system adapts its learning knowledge base 130. Noteworthy, the system adapts concurrently with the use of the system, and not, as in the prior art, as separate and distinct processes.
- We now descri'oe a methodology for generating the tree element of the newly S generated pair.
The first data, or object, field of the tree is set to an encoded identification of the object selected by user 110.
The second data, or advice, field of the tree, which is an encoded statement of procedure that is recommended to be followed when measuring the object identified in 10 the object field, is generated in response to an interaction between user 110 and the troubleshooting expe~t system by way of which user 110 specifies the procedure to measure the object selected by user 110.
The third data, or if-fail, field as well as the fourth data, or if-pass, field are encoded to include empty lists of two-tuple pairs. Inasmuch as both the third and the 15 fourth data fields of the newly generated tree element include empty lists of pairs, that tree is called a leaf.
We now turn to FIG. 2 as an aid in defining a concept called signal path, which will be used in a further descripdon of user acdon classifier 140, user advice classifier 150, and signal path manager 160.
FIG. 2 illustrates a unit 200, which is a generic object, that includes a plurality of objects like objects 221, 224, 229 among others, which are interconnected in some predefined order to provide some funcdon. For purposes of a descripdon of an illustradve embodiment of the principles of our invendon, the particular function is not important and therefore its descripdon is omitted. We assume that a user like user 110 .~: 25 desires to troubleshoot unit 200 and does so in accord with a certain set of rules. Each object as well as the unit itself has an input and an output. For example, input 201 and output 232 are the input and output respectively of unit 200. Input 201 is also the input to object 221 while output 232 is also the output of object 231. Using that background, we now define the concept of a signal path. A signal path is an ordered set of objects 30 from a specific input to a specific output. For example, the signal path from input 241 to output 261 of object 260 includes object 2æ with its output 251, which is also the input 251 to object 250. The output 261 of object 250 is also the output 261 of object 260.
Object 250 has a signal path from its input 251 to its output 261 through both of objects 224 and 225.
g Now the concept of a signal path can be further discussed in terms of a downstream signal path and an upstream signal path.
As mentioned, a signal path is an ordered set of objects from a specific input to a specific output. On the one hand, for a first signal path having an object in the first S signal path, the ~erm downstream signal path means a ~econd signal path consisting of that part of the first signal path, which extends from the output of the object in question and continues to the output of the first signal path. On the other hand, for a first signal path having an object in the first sigDal path, the term upstream signal path means a second signal path consisting of that part of the first signal path, which extends from the 10 input of the first signal path and continues to the input of the object in question. Note that the object in question is itself excluded from the downstream signal path and is also excluded from the upstream signal path. For example, consider a signal path defined from input 241 through objects 260 and 231 to output 232. Now consider object 250. Relative to object 250, the downstream signal path includes object 231 while the upstream signal 15 path includes object 222. As another example, consider a signal path defined from input 201 through object 200 to output 232. Again consider object 250. Relative to object 250, the downstream signal path includes object 231 as before; however, the upstream signal path includes objects 221 and 222. Note that object 270 as well as the plurality of objects illustrated in object 270 are excluded from both the downstream and upstream - ~ 20 signal paths when defined relative to object 250 even though object 270 as well as objects -~ 223, 226, 227, 228, 229 and 230 within object 270 are in the signal path from input 201 through object 200 to output 232.
Now the concept of a signal path can be further discussed in terms of a current signal path.
Consider any signal path with the property that its input is found to be correct while its output is found to be in fault. We can conclude that there is some malfunction in the signal path. The term current signal path means a signal path in which a fault is detected. An encoded identity of the current signal path is stored in a third - memory of the troubleshooting expert system.
It is now convenient to offer a further definition of troubleshooting.
; Troubleshooting can be viewed as the process of idendfying a current signal path. With knowledge of the culTent signal path, it becomes the further task of the troubleshooter to locate the faulty object(s) within the current signal path. The troubleshooter can select an object in the current signal path, measure the output of the selected object, and determine 35 whether or not the output is in fault. On the one hand, if the output is faulty, then either the selected object is at fault or an object in an upstrearn signal path is at fault. On the --~ 2~3~01~
- other hand, if the output of the selected object is not at fault, then an object in the downstream signal path is at fault.
We now return to user action classifier 140 to complete its descripdon. One further rule is em~ in action classifier 140:
Rule 140-3: If the test object selected by user 110 does not belong to the current signal path, then the skill level of user 110 is characterized as that of a beginner troubleshooter.
- The object field and the advice field of the tree element in one of the pairs in the current list of pairs is used: (a) for aiding user 110 in selecting the object to be tested 10 from the current signal path, and (b) for comparing the object selected by user 110 with ' the object identified in the object field of the tree element in the first two-tuple pair in the ' current list of pairs, and (c) for inferring and classifying the action of user 110 in response to the comparison results. For example, the user's actions may be classified according to whether user 110 (i) is or is not using a correct strategy and/or (ii) 15 understands or does not understand the causal dependency between the objects in the current signal path.
That completes our descripdon of user action classifier 140.
The fourth component, user advice classifier 150, could be embodied in software as a set of three rules allowing the troubleshooting expert system to infer and ;~ 20 classify the appropriate kind of advice, which ought to be given user 110, in response to the skill level of user 110.
.
Rule 150-1: If the sldll level of user 110 is that of a beginner ~oubleshooter, then it could be concluded that user 110 needs advice that relates to troubleshooting strategy and, accordingly, advice classifier 150 could so set a flag for later use in 25 signalling the need for that strategy to signal path manager 160.
: Rule 150-2: If the skill level of user 110 is that of a beginner troubleshooter or an expert trouUeshooter, then it could be concluded that user 110 needs advice that relates to the current signal path and, accordingly, advice classifier 150 could so set a fiag for later use in signalling the need for that strategy to signal path manager 160.
; 30 Rule 150-3: If the skill level of user 110 is that of a good troubleshooter, then it would be concluded that user 110 does not need advice and, accordingly, advice ~; classifier 150 could so set a flag for later use in signalling the need for that strategy to signal path manager 160.
That completes our description of user advice classifier 150.
- 203~n~ 1 The fifth component, signal path manager 160, could be embodied in software as a set of two rules allowing the expert system to recursively generate and re-generate the current signal path for the pulpose of seeking out the object in the current signal path, which is at fault.
S Rule 160-1: If the output of the object selected by user 110 is not at fault, then a new culrent signa1 path is generated by the process of (a) retrieving the current signal path from the third memory of the troubleshooting expert system; (b) generating an upstream signal path responsive to the selected object and responsive to the culrent signal path; (c) deleting, or erasing, or removing the selected object and all objects of the 10 upstream signal path from the current signal path to obtain a new current signal path; and (d) storing the new current signal path in the third memory.
Rule 160-2: If the output of the object selected by user 110 is at fault, then anew current signal path is generated by the process of (a) retrieving the cuTrent signal path from the third memory of the troubleshooting expert system; (b) generating a 15 downstream signal path responsive to the selected object and responsive to the current signal path; (c) deleting, or erasing, or removing the selected all objects of the f~, downstream signal path from the current signal path to obtain a new cusrent signal path;
and (d) storing the new current signal path in the third memory.
That completes our description of signal path manager 160.
; 20 The sixth component, graphic output interface 170, could be embodied through a set of two rules allowing the expert system to communicate the appropriate advice to user 110 responsive to the kind of advice identified by the flag set by user advice classifier 150 and responsive to the current list of pairs of the type included in learning knowledge base 130 and responsive to the current signal path.
` 25 Rule 17~1: If the flag set by user advice classifier 150 is set to indicate that . user 110 needs advice that relates to the current signal path, then communicate the current signal path to user 110.
Rule 170-2: If the flag set by user advice classifier 150 is set to indicate that - user 110 needs advice that relates to troubleshooting strategy, then communicate the 30 object field and the advise field from the tree element of the first two-tuple pair in the current list of pairs to user 110.
: Note that the first two-tuple pair includes the tree element with the weight element having the greatest value among the current list of pairs and, according to Rule 140-1, that is the tree which will be used most often and therefore is in a sense the best 35 troubleshoodng advice.
~,~
2~3 ~ 011 ;; Note also that pursuant to Rules 170-1 and 170-2, the only advice - ~ communicated to user 110 is advice, which is relevant to the kind of troubleshooting error actually detected by user action classifier 140. Moreover, as user 110 is not required , to follow the advice .lt each and every step of the troubleshooting process, there is less S time wasted by the consultation module, i.e. unproductive time is remarkably decreased.
Although our invention has been described and illustrated in detail, it is to beunderstood that the same is not by way of limitation. Hence, the spirit and scope of our invention is limited only by the terms of the appended claims.
.
., .
.
',' ' .
:.
, .
Claims (12)
1. A method for troubleshooting an object in a system under test, the method including the steps of interactively communicating between a user and a troubleshooting system, generating a learning knowledge base, identifying an object being tested by the user, utilizing the learning knowledge base for troubleshooting the test object, and classifying the test object as faulty or not faulty AND CHARACTERIZED IN THAT the method comprises the steps of:
(a) structuring prior actions of the user as pairs of elements in a multi-branch tree in the learning knowledge base;
(b) responsive to selected pairs of elements, comparing present actions of the user with the prior actions of the user;
(c) generating a current signal path between a first input and a first output inthe system under test;
(d) determining whether the test object belongs to the current signal path;
(e) responsive to the comparison of the actions of the user and responsive to whether the test object belongs to the current path, classifying the skill level of the user into one of a plurality of skill levels;
(f) responsive to the skill level classification of the user, classifying the kind of troubleshooting advice to be given the user as to one of a plurality of kinds of troubleshooting advice; and (g) communicating the classified troubleshooting advice to the user.
(a) structuring prior actions of the user as pairs of elements in a multi-branch tree in the learning knowledge base;
(b) responsive to selected pairs of elements, comparing present actions of the user with the prior actions of the user;
(c) generating a current signal path between a first input and a first output inthe system under test;
(d) determining whether the test object belongs to the current signal path;
(e) responsive to the comparison of the actions of the user and responsive to whether the test object belongs to the current path, classifying the skill level of the user into one of a plurality of skill levels;
(f) responsive to the skill level classification of the user, classifying the kind of troubleshooting advice to be given the user as to one of a plurality of kinds of troubleshooting advice; and (g) communicating the classified troubleshooting advice to the user.
2. The method defined in claim 1 CHARACTERIZED IN THAT
the method further comprises the steps of:
(h) measuring the output of the test object and (i) responsive to the output of the test object, generating a new current signal path.
the method further comprises the steps of:
(h) measuring the output of the test object and (i) responsive to the output of the test object, generating a new current signal path.
3. The method defined in claim 2 CHARACTERIZED IN THAT
the method further comprises the steps of:
(j) responsive to the output of the test object evidencing a fault in the test object, generating a downstream signal path as the new current signal path.
the method further comprises the steps of:
(j) responsive to the output of the test object evidencing a fault in the test object, generating a downstream signal path as the new current signal path.
4. The method defined in claim 2 CHARACTERIZED IN THAT
the method further comprises the steps of:
(j) responsive to the output of the test object evidencing no fault in the test object, generating an upstream signal path as the new current signal path.
the method further comprises the steps of:
(j) responsive to the output of the test object evidencing no fault in the test object, generating an upstream signal path as the new current signal path.
5. A system for troubleshooting an object under test, the system including means for interactively communicating between a user and a troubleshooting system, means for generating a learning knowledge base, means for identifying an object being tested by the user, means for utilizing the learning knowledge base for troubleshooting the test object, and means for classifying the test object as faulty or not faulty and CHARACTERIZED IN THAT
the system comprises:
means for structuring prior actions of the user as pairs in a multi-branch tree in the learning knowledge base;
means, responsive to selected pairs of elements, for comparing present actions of the user with the prior actions of the user;
means for generating a current signal path between a first input and a first output in the system under test;
means for determining whether the test object belongs to the current signal path;
means responsive to the comparison of the actions of the user and responsive to whether the test object belongs to the current path for classifying the skill level of the user into one of a plurality of skill levels;
means responsive to the skill level classification of the user for classifying the kind of troubleshooting advice to be given the user as to one of a plurality of kinds of troubleshooting advice; and means for communicating the classified troubleshooting advice to the user.
the system comprises:
means for structuring prior actions of the user as pairs in a multi-branch tree in the learning knowledge base;
means, responsive to selected pairs of elements, for comparing present actions of the user with the prior actions of the user;
means for generating a current signal path between a first input and a first output in the system under test;
means for determining whether the test object belongs to the current signal path;
means responsive to the comparison of the actions of the user and responsive to whether the test object belongs to the current path for classifying the skill level of the user into one of a plurality of skill levels;
means responsive to the skill level classification of the user for classifying the kind of troubleshooting advice to be given the user as to one of a plurality of kinds of troubleshooting advice; and means for communicating the classified troubleshooting advice to the user.
6. The system defined in claim 5 CHARACTERIZED IN THAT
the system further comprises:
means for measuring the output of the test object and means responsive to the output of the test object for generating a new current signal path.
the system further comprises:
means for measuring the output of the test object and means responsive to the output of the test object for generating a new current signal path.
7. The system defined in claim 6 CHARACTERIZED IN THAT
the system further comprises:
means responsive to the output of the test object evidencing a fault in the testobject for generating a downstream signal path as the new current signal path.
the system further comprises:
means responsive to the output of the test object evidencing a fault in the testobject for generating a downstream signal path as the new current signal path.
8. The system defined in claim 6 CHARACTERIZED IN THAT
the system further comprises:
means responsive to the output of the test object evidencing no fault in the test object for generating an upstream signal path as the new current signal path.
the system further comprises:
means responsive to the output of the test object evidencing no fault in the test object for generating an upstream signal path as the new current signal path.
9. Apparatus for troubleshooting an object in a system under test, the apparatus including means for interactively communicating between a user and a troubleshooting system, means for generating a learning knowledge base, means for identifying an object being tested by the user, means for utilizing the learning knowledge base for troubleshooting the test object, and means for classifying the test object as faulty or not faulty and CHARACTERIZED IN THAT
the apparatus comprises:
means for structuring prior actions of the user as pairs in a multi-branch tree in the learning knowledge base;
means, responsive to selected pairs of elements, for comparing present actions of the user with the prior actions of the user;
means for generating a current signal path between a first input and a first output in the system under test;
means for determining whether the test object belongs to the current signal path;
means responsive to the comparison of the actions of the user and responsive to whether the test object belongs to the current path for classifying the skill level of the user into one of a plurality of skill levels;
means responsive to the skill level classification of the user for classifying the kind of troubleshooting advice to be given the user as to one of a plurality of kinds of troubleshooting advice; and means for communicating the classified troubleshooting advice to the user.
the apparatus comprises:
means for structuring prior actions of the user as pairs in a multi-branch tree in the learning knowledge base;
means, responsive to selected pairs of elements, for comparing present actions of the user with the prior actions of the user;
means for generating a current signal path between a first input and a first output in the system under test;
means for determining whether the test object belongs to the current signal path;
means responsive to the comparison of the actions of the user and responsive to whether the test object belongs to the current path for classifying the skill level of the user into one of a plurality of skill levels;
means responsive to the skill level classification of the user for classifying the kind of troubleshooting advice to be given the user as to one of a plurality of kinds of troubleshooting advice; and means for communicating the classified troubleshooting advice to the user.
10. The apparatus defined in claim 9 and CHARACTERIZED IN THAT
the apparatus further comprises:
means for measuring the output of the test object and means responsive to the output of the test object for generating a new current signal path.
the apparatus further comprises:
means for measuring the output of the test object and means responsive to the output of the test object for generating a new current signal path.
11. The system defined in claim 10 CHARACTERIZED IN THAT
the system further comprises:
means responsive to the output of the test object evidencing a fault in the testobject for generating a downstream signal path as the new current signal path.
the system further comprises:
means responsive to the output of the test object evidencing a fault in the testobject for generating a downstream signal path as the new current signal path.
12. The system defined in claim 10 CHARACTERIZED IN THAT
the system further comprises:
means responsive to the output of the test object evidencing no fault in the test object for generating an upstream signal path as the new current signal path.
the system further comprises:
means responsive to the output of the test object evidencing no fault in the test object for generating an upstream signal path as the new current signal path.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US516,411 | 1990-04-30 | ||
| US07/516,411 US5107499A (en) | 1990-04-30 | 1990-04-30 | Arrangement for automated troubleshooting using selective advice and a learning knowledge base |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CA2037011A1 CA2037011A1 (en) | 1991-10-31 |
| CA2037011C true CA2037011C (en) | 1994-05-31 |
Family
ID=24055459
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA002037011A Expired - Fee Related CA2037011C (en) | 1990-04-30 | 1991-02-25 | Arrangement for automated troubleshooting using selective advice and a learning knowledge base |
Country Status (5)
| Country | Link |
|---|---|
| US (1) | US5107499A (en) |
| EP (1) | EP0456378A3 (en) |
| JP (1) | JPH04229328A (en) |
| KR (1) | KR950012221B1 (en) |
| CA (1) | CA2037011C (en) |
Families Citing this family (49)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| SE467026B (en) * | 1990-09-21 | 1992-05-11 | Televerket | DEVICE FOR STRUCTURING TECHNICAL INFORMATION WHEN CREATING A KNOWLEDGE DATABASE AND TROUBLESHOOTING IN TECHNICAL EQUIPMENT |
| US5195095A (en) * | 1990-12-28 | 1993-03-16 | General Electric Company | Algorithm for identifying tests to perform for fault isolation |
| US5418889A (en) * | 1991-12-02 | 1995-05-23 | Ricoh Company, Ltd. | System for generating knowledge base in which sets of common causal relation knowledge are generated |
| US5297150A (en) * | 1992-06-17 | 1994-03-22 | International Business Machines Corporation | Rule-based method for testing of programming segments |
| US5313241A (en) * | 1992-11-25 | 1994-05-17 | Eastman Kodak Company | Processor diagnostics using switch settings |
| US5546507A (en) * | 1993-08-20 | 1996-08-13 | Unisys Corporation | Apparatus and method for generating a knowledge base |
| US5473732A (en) * | 1993-11-02 | 1995-12-05 | Chang; Hou-Mei H. | Relational artificial intelligence system |
| US6338148B1 (en) | 1993-11-10 | 2002-01-08 | Compaq Computer Corporation | Real-time test controller |
| US5490149A (en) * | 1993-12-28 | 1996-02-06 | International Business Machines Corporation | Tactical read error recovery strategy based on dynamic feedback |
| CA2148028A1 (en) * | 1994-05-25 | 1995-11-26 | Deborah L. Mcguinness | Knowledge base management system with dependency information for procedural tests |
| US5544308A (en) * | 1994-08-02 | 1996-08-06 | Giordano Automation Corp. | Method for automating the development and execution of diagnostic reasoning software in products and processes |
| JPH08137647A (en) * | 1994-11-07 | 1996-05-31 | Fujitsu Ltd | Method and device for reproducing operation guide information reproduction, method and device for reproducing multimedia information |
| US5787232A (en) * | 1996-01-10 | 1998-07-28 | Siemens Corporate Research, Inc. | Efficient data-driven theory revision system |
| JP3195226B2 (en) * | 1996-03-14 | 2001-08-06 | 株式会社東芝 | Business support system |
| US6728343B1 (en) | 1998-03-16 | 2004-04-27 | Conair Corporation | User appliance with voice help guide |
| US6604092B1 (en) | 1999-02-26 | 2003-08-05 | Lisa E. Stewart | Expert system utilizing a knowledge base and design of experiment (DOE) techniques |
| JP2003527184A (en) * | 2000-03-23 | 2003-09-16 | コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ | Remote medical imaging system diagnosis |
| US6829734B1 (en) | 2000-04-04 | 2004-12-07 | International Business Machines Corporation | Method for discovering problem resolutions in a free form computer helpdesk data set |
| AU5558401A (en) | 2000-04-20 | 2001-11-07 | General Electric Company | Method and system for graphically identifying replacement parts for generally complex equipment |
| WO2002017273A2 (en) * | 2000-08-23 | 2002-02-28 | General Electric Company | Method for training service personnel to service selected equipment |
| US7016809B1 (en) * | 2000-09-26 | 2006-03-21 | Intel Corporation | Method and system for automatically interpreting computer system performance measurements |
| MXPA02008345A (en) | 2000-12-29 | 2002-12-13 | Gen Electric | Method and system for identifying repeatedly malfunctioning equipment. |
| US20020156692A1 (en) * | 2001-04-20 | 2002-10-24 | Squeglia Mark R. | Method and system for managing supply of replacement parts of a piece of equipment |
| US6895573B2 (en) | 2001-10-26 | 2005-05-17 | Resultmaker A/S | Method for generating a workflow on a computer, and a computer system adapted for performing the method |
| US6882271B2 (en) | 2002-11-12 | 2005-04-19 | Koninklijke Philips Electronics N.V. | Automated medical imaging system repair diagnostics |
| US6865512B2 (en) | 2002-11-12 | 2005-03-08 | Koninklijke Philips Electronics N.V. | Automated medical imaging system maintenance diagnostics |
| US7500145B2 (en) * | 2004-05-28 | 2009-03-03 | International Business Machines Corporation | Anomaly-driven software switch to capture event responses and automate recovery |
| US20060156141A1 (en) * | 2004-12-07 | 2006-07-13 | Ouchi Norman K | Defect symptom repair system and methods |
| US8326925B2 (en) * | 2005-01-31 | 2012-12-04 | Hewlett-Packard Development Company, L.P. | Connecting to experts in a discussion board |
| US20070157311A1 (en) * | 2005-12-29 | 2007-07-05 | Microsoft Corporation | Security modeling and the application life cycle |
| US20070192344A1 (en) * | 2005-12-29 | 2007-08-16 | Microsoft Corporation | Threats and countermeasures schema |
| US7890315B2 (en) | 2005-12-29 | 2011-02-15 | Microsoft Corporation | Performance engineering and the application life cycle |
| US20070156420A1 (en) * | 2005-12-29 | 2007-07-05 | Microsoft Corporation | Performance modeling and the application life cycle |
| US7818788B2 (en) * | 2006-02-14 | 2010-10-19 | Microsoft Corporation | Web application security frame |
| US7712137B2 (en) | 2006-02-27 | 2010-05-04 | Microsoft Corporation | Configuring and organizing server security information |
| US8667332B2 (en) * | 2007-04-24 | 2014-03-04 | Honeywell International Inc. | Methods for optimizing diagnostics reasoner models |
| US20090138854A1 (en) * | 2007-11-22 | 2009-05-28 | Suresh Mani | Software error detection method, software module, database and system |
| FR2938675B1 (en) * | 2008-11-14 | 2010-11-12 | Thales Sa | DEVICE FOR IMPROVING MAINTENANCE PROCEDURES FOR ON-BOARD EQUIPMENT |
| DE102009004446A1 (en) * | 2009-01-13 | 2010-07-15 | Siemens Aktiengesellschaft | Sub-entity's operating actions e.g. luggage transportation system's aircraft service works, coordinating method for organization i.e. airport, involves producing new rules based on operating actions of sub-entities and condition data |
| US8751879B2 (en) * | 2011-03-30 | 2014-06-10 | International Business Machines Corporation | Intelligently monitoring and dispatching information technology service alerts |
| RU2460122C1 (en) * | 2011-04-19 | 2012-08-27 | Закрытое акционерное общество "Лаборатория Касперского" | System and method of raising security level of computer system |
| US9239991B2 (en) | 2013-09-05 | 2016-01-19 | General Electric Company | Services support system and method |
| US10339681B2 (en) | 2013-11-22 | 2019-07-02 | Raytheon Company | Interactive multimedia process flow chart builder |
| EP3059676B1 (en) * | 2015-02-20 | 2019-09-11 | Siemens Aktiengesellschaft | A method and apparatus for analyzing the availability of a system, in particular of a safety critical system |
| US10073753B2 (en) * | 2016-02-14 | 2018-09-11 | Dell Products, Lp | System and method to assess information handling system health and resource utilization |
| US10908983B2 (en) * | 2017-03-31 | 2021-02-02 | Cae Inc. | Method and system for preventing an anomaly in a simulator |
| US11468235B2 (en) | 2019-01-21 | 2022-10-11 | Bayestree Intelligence Pvt Ltd. | Recommender and remediation system for enterprise service management |
| CN113159993B (en) * | 2021-04-07 | 2022-08-12 | 国家电网有限公司客户服务中心 | Knowledge base construction method and system based on power system |
| KR102381417B1 (en) * | 2021-10-15 | 2022-04-01 | 로봇앤컴주식회사 | Devices, systems, methods and programs providing one-stop online repair service |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4864492A (en) * | 1986-09-17 | 1989-09-05 | International Business Machines Corporation | System and method for network configuration |
| WO1988005918A1 (en) * | 1987-02-06 | 1988-08-11 | Analytics Incorporated | Maintenance system |
| US4965741A (en) * | 1988-10-17 | 1990-10-23 | Ncr Corporation | Method for providing an improved human user interface to a knowledge based system |
-
1990
- 1990-04-30 US US07/516,411 patent/US5107499A/en not_active Expired - Fee Related
-
1991
- 1991-02-25 CA CA002037011A patent/CA2037011C/en not_active Expired - Fee Related
- 1991-04-22 JP JP3116616A patent/JPH04229328A/en active Pending
- 1991-04-23 KR KR1019910006477A patent/KR950012221B1/en not_active IP Right Cessation
- 1991-04-24 EP EP19910303706 patent/EP0456378A3/en not_active Withdrawn
Also Published As
| Publication number | Publication date |
|---|---|
| CA2037011A1 (en) | 1991-10-31 |
| US5107499A (en) | 1992-04-21 |
| EP0456378A3 (en) | 1992-12-23 |
| KR950012221B1 (en) | 1995-10-16 |
| KR910018908A (en) | 1991-11-30 |
| JPH04229328A (en) | 1992-08-18 |
| EP0456378A2 (en) | 1991-11-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA2037011C (en) | Arrangement for automated troubleshooting using selective advice and a learning knowledge base | |
| KR930004437B1 (en) | Technique for producing an expert system for system fault diagnosis | |
| Chen et al. | A choice relation framework for supporting category-partition test case generation | |
| Ng | Model-based, multiple fault diagnosis of time-varying, continuous physical devices | |
| Henry et al. | A standard interface for self-validating sensors | |
| Menzies | Practical machine learning for software engineering and knowledge engineering | |
| Nursimulu et al. | Cause-effect graphing analysis and validation of requirements | |
| Park et al. | Fusion-based sensor fault detection | |
| Rask et al. | Simulation and comparison of Albrecht's function point and DeMarco's function bang metrics in a CASE environment | |
| Friedland | Maximum-likelihood estimation of a process with random transitions (failures) | |
| Havens et al. | Platypus: A Constraint-Based Reasoning System. | |
| Swamy et al. | Cointegration: Its fatal flaw and a proposed solution | |
| Kahn | From Application Shell to Knowledge Acquisition System. | |
| Hegedüs | Combinatorial results on the complexity of teaching and learning | |
| Singh | Saturn: An Automatic Test Generation System for Digital Circuits. | |
| Bahi et al. | Basins of attraction in fully asynchronous discrete-time discrete-state dynamic networks | |
| Stefanowicz | Selected issues of data editing | |
| Grastien et al. | Diagnosability of Discrete Faults with Uncertain Observations | |
| JPH04273540A (en) | Fault diagnosing method and device | |
| Thistle et al. | Analysis of arbitrarily large networks of discrete-event systems | |
| Åkerlund | Model-based diagnosis using MathModelica | |
| Ashley et al. | Diagnosis of condition systems using causal structure | |
| Wegener | Test synthesis: an innovative approach to analog and system testability analysis using state dependent flow diagrams | |
| Hu | Sensitivity of Graph-theoretic Metrics to Edge Directions for Structured and Unstructured Programs� | |
| Yuan | Modeling software quality with TREEDISC algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| EEER | Examination request | ||
| MKLA | Lapsed |